{ "schemaVersion": 1, "operationId": "20260831040323-d1ccc05be5", "checkedAt": "2026-08-31T04:20:19.954Z", "status": "pinned-range-read-glm52-sglang-moe-bias-fp32-audit", "sourceReceipts": [ { "id": "zai-glm52-release", "url": "https://z.ai/blog/glm-5.2", "finalUrl": "https://z.ai/blog/glm-5.2", "purpose": "mandatory first-party GLM-5.2 release check", "httpStatus": 200, "bytes": 598, "sha256": "a9e8c2b6f34717d69e3a0aa26bb117256a4d8c95bd299910c2a693000ee88fe8" }, { "id": "zhipu-research-index", "url": "http://zhipuai.cn/zh/research", "finalUrl": "https://www.zhipuai.cn/zh/research", "purpose": "mandatory Zhipu AI research-discovery check", "httpStatus": 200, "bytes": 1235119, "sha256": "7d50f4c290fbc240f50fabc8d18f7a499688f968f53951655572909ddf547d4e" }, { "id": "sglang-pr-37133", "url": "https://api.github.com/repos/sgl-project/sglang/pulls/37133", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/pulls/37133", "purpose": "capture the proposed GLM-5.2 fp32 correction-bias fix and author-reported observations", "httpStatus": 200, "bytes": 28052, "sha256": "0260d551f9e5d585ecd9a206dc308ab5f1eb49130b161f981fc7baaf27f52d26" }, { "id": "sglang-pr-37133-files", "url": "https://api.github.com/repos/sgl-project/sglang/pulls/37133/files?per_page=100", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/pulls/37133/files?per_page=100", "purpose": "capture the exact three-file proposal surface", "httpStatus": 200, "bytes": 10565, "sha256": "dfc66372d12b279efc137ce978857c723942b2dd1c9b399eff168b2a41603f59" }, { "id": "sglang-pr-37133-commits", "url": "https://api.github.com/repos/sgl-project/sglang/pulls/37133/commits?per_page=100", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/pulls/37133/commits?per_page=100", "purpose": "pin the proposal commit", "httpStatus": 200, "bytes": 5403, "sha256": "652d002b3c7cf8f38218ba94f67ed94c76301b3dec44d7502550f24e49ae460d" }, { "id": "sglang-pr-37133-reviews", "url": "https://api.github.com/repos/sgl-project/sglang/pulls/37133/reviews?per_page=100", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/pulls/37133/reviews?per_page=100", "purpose": "capture review state", "httpStatus": 200, "bytes": 2, "sha256": "4f53cda18c2baa0c0354bb5f9a3ecbe5ed12ab4d8e11ba873c2f11161202b945" }, { "id": "sglang-pr-37133-comments", "url": "https://api.github.com/repos/sgl-project/sglang/issues/37133/comments?per_page=100", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/issues/37133/comments?per_page=100", "purpose": "capture maintainer discussion state", "httpStatus": 200, "bytes": 2, "sha256": "4f53cda18c2baa0c0354bb5f9a3ecbe5ed12ab4d8e11ba873c2f11161202b945" }, { "id": "sglang-pr-base-run", "url": "https://api.github.com/repos/sgl-project/sglang/actions/runs/33312609040", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/actions/runs/33312609040", "purpose": "capture the aggregate base-test workflow state named in the PR", "httpStatus": 200, "bytes": 16516, "sha256": "bd0e734b60fe5304264325b9df90613c7c979ca80beb728c92068d751dfcc0b1" }, { "id": "sglang-pr-extra-run", "url": "https://api.github.com/repos/sgl-project/sglang/actions/runs/33312608877", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/actions/runs/33312608877", "purpose": "capture the aggregate extra-test workflow state named in the PR", "httpStatus": 200, "bytes": 15496, "sha256": "e644a5fe025faa8204fbd2088c119f408ed70d4028deaed9bb29eb7ccf0d9001" }, { "id": "sglang-pr-amd-run", "url": "https://api.github.com/repos/sgl-project/sglang/actions/runs/33312608974", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/actions/runs/33312608974", "purpose": "capture the aggregate AMD workflow state named in the PR", "httpStatus": 200, "bytes": 14925, "sha256": "c2cf28781f7d320a8d4049452aa1ce5ba574b6e948d13f2aa0641ec6e157ef38" }, { "id": "sglang-base-deepseek-v2", "url": "https://raw.githubusercontent.com/sgl-project/sglang/26c754e06ea60c2293098a1faa75f622c8256ab0/python/sglang/srt/models/deepseek_v2.py", "finalUrl": "https://raw.githubusercontent.com/sgl-project/sglang/26c754e06ea60c2293098a1faa75f622c8256ab0/python/sglang/srt/models/deepseek_v2.py", "purpose": "verify the base parameter-construction downcast", "httpStatus": 200, "bytes": 135323, "sha256": "d6339ecde42f4cc1135f4d278b53a1d08c06f32e035b2adafe0e5b799503934e" }, { "id": "sglang-head-deepseek-v2", "url": "https://raw.githubusercontent.com/sgl-project/sglang/60df2db3e83dff36cbde1f7632c76b5d3864cb9a/python/sglang/srt/models/deepseek_v2.py", "finalUrl": "https://raw.githubusercontent.com/sgl-project/sglang/60df2db3e83dff36cbde1f7632c76b5d3864cb9a/python/sglang/srt/models/deepseek_v2.py", "purpose": "pin the proposed GlmMoeDsa architecture gate", "httpStatus": 200, "bytes": 135883, "sha256": "11bfd339b299e2db1d3c97ca118c2bdce99bc9511811b55f03e5095537a7a80e" }, { "id": "sglang-base-topk", "url": "https://raw.githubusercontent.com/sgl-project/sglang/26c754e06ea60c2293098a1faa75f622c8256ab0/python/sglang/srt/layers/moe/topk.py", "finalUrl": "https://raw.githubusercontent.com/sgl-project/sglang/26c754e06ea60c2293098a1faa75f622c8256ab0/python/sglang/srt/layers/moe/topk.py", "purpose": "verify the base AITER boundary downcast", "httpStatus": 200, "bytes": 101086, "sha256": "ff34c4f381af401e250af8fc2a94aa1c00051dfbd40b89a78a77d62ae451213c" }, { "id": "sglang-head-topk", "url": "https://raw.githubusercontent.com/sgl-project/sglang/60df2db3e83dff36cbde1f7632c76b5d3864cb9a/python/sglang/srt/layers/moe/topk.py", "finalUrl": "https://raw.githubusercontent.com/sgl-project/sglang/60df2db3e83dff36cbde1f7632c76b5d3864cb9a/python/sglang/srt/layers/moe/topk.py", "purpose": "pin the proposed AITER fp32 bias and logit path", "httpStatus": 200, "bytes": 101706, "sha256": "19a456d6250bc3bc8d43975a31a1c8cf04da1d0cdfe3d255e900bc894aa86013" }, { "id": "sglang-head-tests", "url": "https://raw.githubusercontent.com/sgl-project/sglang/60df2db3e83dff36cbde1f7632c76b5d3864cb9a/test/registered/unit/models/test_glmmoedsa_correction_bias_fp32.py", "finalUrl": "https://raw.githubusercontent.com/sgl-project/sglang/60df2db3e83dff36cbde1f7632c76b5d3864cb9a/test/registered/unit/models/test_glmmoedsa_correction_bias_fp32.py", "purpose": "pin the nine proposed CPU-only regression tests", "httpStatus": 200, "bytes": 6040, "sha256": "1c5fdcb504c8369262bac5a67dace8a70f1fb683b3f4a441fd4232b8f5400017" }, { "id": "sglang-head-model-config", "url": "https://raw.githubusercontent.com/sgl-project/sglang/60df2db3e83dff36cbde1f7632c76b5d3864cb9a/python/sglang/srt/configs/model_config.py", "finalUrl": "https://raw.githubusercontent.com/sgl-project/sglang/60df2db3e83dff36cbde1f7632c76b5d3864cb9a/python/sglang/srt/configs/model_config.py", "purpose": "verify the GLM NextN architecture rewrite covered by the proposal", "httpStatus": 200, "bytes": 97296, "sha256": "2a229f3afee04de0df4768ac1c58175d9cba9844a50acc07647916a496d1f566" }, { "id": "sglang-latest-release", "url": "https://api.github.com/repos/sgl-project/sglang/releases/latest", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/releases/latest", "purpose": "separate the latest tagged release from an open proposal", "httpStatus": 200, "bytes": 54159, "sha256": "f026d905aaecb7c78354275630fa52bd3704e2eba60c656b7c09faf48370c466" }, { "id": "sglang-v0518-commit", "url": "https://api.github.com/repos/sgl-project/sglang/commits/v0.5.18", "finalUrl": "https://api.github.com/repos/sgl-project/sglang/commits/v0.5.18", "purpose": "pin the most recent tagged release at audit time", "httpStatus": 200, "bytes": 12417, "sha256": "8241eec1d54eea1101e9c713c396de5c10fbed9fa30bc44895545deb6f1f4ff8" }, { "id": "hf-model-config", "url": "https://huggingface.co/zai-org/GLM-5.2-FP8/resolve/ba978f7d347eaf65d22f1a86833408afdb953541/config.json", "finalUrl": "https://huggingface.co/api/resolve-cache/models/zai-org/GLM-5.2-FP8/ba978f7d347eaf65d22f1a86833408afdb953541/config.json?%2Fzai-org%2FGLM-5.2-FP8%2Fresolve%2Fba978f7d347eaf65d22f1a86833408afdb953541%2Fconfig.json=&etag=%224e1f0168afd127189fb1c4ddb1d4476a4fca96ac%22", "purpose": "pin the official GLM-5.2 FP8 MoE routing and quantization contract", "httpStatus": 200, "bytes": 29464, "sha256": "22e49334abf8562fecf70ca3292ba3f5b33f5602fb2bf10b52dd64a66cfe65ff" }, { "id": "hf-model-index", "url": "https://huggingface.co/zai-org/GLM-5.2-FP8/resolve/ba978f7d347eaf65d22f1a86833408afdb953541/model.safetensors.index.json", "finalUrl": "https://huggingface.co/zai-org/GLM-5.2-FP8/resolve/ba978f7d347eaf65d22f1a86833408afdb953541/model.safetensors.index.json", "purpose": "map every correction-bias tensor to its immutable shard", "httpStatus": 200, "bytes": 11359251, "sha256": "e0fe7f28c1f853d4824e4d796374e3dacf1fe470988773952c79b063768134bf" }, { "id": "hf-model-card", "url": "https://huggingface.co/zai-org/GLM-5.2-FP8/resolve/ba978f7d347eaf65d22f1a86833408afdb953541/README.md", "finalUrl": "https://huggingface.co/api/resolve-cache/models/zai-org/GLM-5.2-FP8/ba978f7d347eaf65d22f1a86833408afdb953541/README.md?%2Fzai-org%2FGLM-5.2-FP8%2Fresolve%2Fba978f7d347eaf65d22f1a86833408afdb953541%2FREADME.md=&etag=%2242c3f8e6cad8e812f71c352381eaf058ff91a217%22", "purpose": "capture the official checkpoint card and attribution", "httpStatus": 200, "bytes": 10909, "sha256": "de23c1b7cab43a99f0fedf4edf10de0d165882a2ecb2e2bad6c5796fcabf2e46" }, { "id": "hf-model-revision-metadata", "url": "https://huggingface.co/api/models/zai-org/GLM-5.2-FP8/revision/ba978f7d347eaf65d22f1a86833408afdb953541?expand=safetensors", "finalUrl": "https://huggingface.co/api/models/zai-org/GLM-5.2-FP8/revision/ba978f7d347eaf65d22f1a86833408afdb953541?expand=safetensors", "purpose": "cross-check pinned parameter dtype totals", "httpStatus": 200, "bytes": 169, "sha256": "ca7f30bc8a5809d49eade50361b631311b08c9371e56bec7217df8c4d1f324fd" } ], "pins": { "sglangPrNumber": 37133, "sglangBaseRevision": "26c754e06ea60c2293098a1faa75f622c8256ab0", "sglangHeadRevision": "60df2db3e83dff36cbde1f7632c76b5d3864cb9a", "modelId": "zai-org/GLM-5.2-FP8", "modelRevision": "ba978f7d347eaf65d22f1a86833408afdb953541" }, "proposal": { "number": 37133, "title": "[GLM-5.2] Keep GlmMoeDsa MoE e_score_correction_bias in fp32", "url": "https://github.com/sgl-project/sglang/pull/37133", "state": "open", "draft": false, "merged": false, "mergedAt": null, "mergeableState": "unstable", "createdAt": "2026-08-30T12:51:24Z", "updatedAt": "2026-08-30T12:52:07Z", "headSha": "60df2db3e83dff36cbde1f7632c76b5d3864cb9a", "baseSha": "26c754e06ea60c2293098a1faa75f622c8256ab0", "commitCount": 1, "changedFiles": [ "python/sglang/srt/layers/moe/topk.py", "python/sglang/srt/models/deepseek_v2.py", "test/registered/unit/models/test_glmmoedsa_correction_bias_fp32.py" ], "additions": 164, "deletions": 3, "approvalCount": 0, "reviewCount": 0, "discussionCount": 0, "authorReported": { "biasRange": [ 6.817, 7.063 ], "spread": 0.246, "fp32Distinct": 238, "bf16Distinct": 8, "differentTop8TokenPercent": 98.5, "gsm8kBefore": 0.941, "gsm8kAfter": 0.947, "gpqaDiamondBefore": 0.8333, "gpqaDiamondAfter": 0.8182, "microsecondsPerGatingCall": 4.5, "maximumDecodeStepPercent": 1.2, "bodyContainsAllValues": true, "evidenceOwner": "SGLang PR author; not reproduced by this audit" } }, "workflowState": [ { "id": 33312609040, "name": "PR #37133 - [GLM-5.2] Keep GlmMoeDsa MoE e_score_correction_bias in fp32", "event": "pull_request", "status": "completed", "conclusion": "failure", "createdAt": "2026-08-30T12:51:29Z", "updatedAt": "2026-08-30T12:52:03Z", "headSha": "60df2db3e83dff36cbde1f7632c76b5d3864cb9a", "url": "https://github.com/sgl-project/sglang/actions/runs/33312609040" }, { "id": 33312608877, "name": "PR Test Extra", "event": "pull_request", "status": "completed", "conclusion": "failure", "createdAt": "2026-08-30T12:51:29Z", "updatedAt": "2026-08-30T12:52:06Z", "headSha": "60df2db3e83dff36cbde1f7632c76b5d3864cb9a", "url": "https://github.com/sgl-project/sglang/actions/runs/33312608877" }, { "id": 33312608974, "name": "PR Test ROCm 7.2 (AMD)", "event": "pull_request", "status": "completed", "conclusion": "failure", "createdAt": "2026-08-30T12:51:29Z", "updatedAt": "2026-08-30T12:51:59Z", "headSha": "60df2db3e83dff36cbde1f7632c76b5d3864cb9a", "url": "https://github.com/sgl-project/sglang/actions/runs/33312608974" } ], "sourceBoundary": { "baseParameterSiteDowncastsAiterQuantizedBias": true, "headArchitectureHelperPresent": true, "headParameterGatePresent": true, "baseAiterBoundaryDowncastsBias": true, "headAiterFp32BranchPresent": true, "headAiterLegacyBranchPresent": true, "cpuTestCount": 9, "testsRegisterCpuCi": true, "testsCoverMainAndNextN": true, "testsKeepNonGlmBf16": true, "testsUseSyntheticThirtyFourFixture": true, "nextNRewritePresent": true }, "releaseBoundary": { "latestReleaseTag": "v0.5.18", "latestReleasePublishedAt": "2026-08-22T00:09:15Z", "v0518ResolvedSha": "71de97b264b04dcd514cf904003028aefe9775c8", "v0518CommittedAt": "2026-08-20T21:29:24Z", "proposalIncludedInLatestTag": false }, "modelContract": { "architectures": [ "GlmMoeDsaForCausalLM" ], "modelType": "glm_moe_dsa", "numHiddenLayers": 78, "numNextnPredictLayers": 1, "routedExperts": 256, "expertsPerToken": 8, "topkMethod": "noaux_tc", "routedScalingFactor": 2.5, "firstDenseLayers": 3, "moeLayerFrequency": 1, "quantizationMethod": "fp8", "modulesToNotConvert": [ "model.layers.59.post_attention_layernorm", "model.layers.47.mlp.gate.e_score_correction_bias", "model.layers.71.self_attn.kv_a_layernorm", "model.layers.46.mlp.gate.e_score_correction_bias", "model.layers.30.self_attn.kv_a_layernorm", "model.layers.23.self_attn.kv_a_layernorm", "model.layers.73.mlp.gate.e_score_correction_bias", "model.layers.36.mlp.gate", "model.layers.12.self_attn.q_a_layernorm", "model.layers.51.mlp.gate", "model.layers.47.self_attn.kv_a_layernorm", "model.layers.45.self_attn.q_a_layernorm", "model.layers.15.self_attn.q_a_layernorm", "model.layers.39.input_layernorm", "model.layers.50.self_attn.kv_a_layernorm", "model.layers.19.post_attention_layernorm", "model.layers.71.input_layernorm", "model.layers.72.self_attn.kv_a_layernorm", "model.layers.70.mlp.gate", "model.layers.52.input_layernorm", "model.layers.78.self_attn.kv_a_layernorm", "model.layers.48.post_attention_layernorm", "model.layers.4.self_attn.q_a_layernorm", "model.layers.62.self_attn.q_a_layernorm", "model.layers.4.post_attention_layernorm", "model.layers.38.self_attn.indexers_proj", "model.layers.61.mlp.gate.e_score_correction_bias", "model.layers.14.mlp.gate.e_score_correction_bias", "model.layers.36.self_attn.kv_a_layernorm", "model.layers.65.mlp.gate", "model.layers.13.self_attn.kv_a_layernorm", "model.layers.57.mlp.gate", "model.layers.65.self_attn.kv_a_layernorm", "model.layers.68.input_layernorm", "model.layers.70.self_attn.kv_a_layernorm", "model.layers.24.mlp.gate.e_score_correction_bias", "model.layers.4.input_layernorm", "model.layers.20.input_layernorm", "model.layers.65.self_attn.q_a_layernorm", "model.layers.38.mlp.gate", "model.layers.76.mlp.gate.e_score_correction_bias", "model.layers.62.input_layernorm", "model.layers.45.input_layernorm", "model.layers.68.post_attention_layernorm", "model.layers.72.input_layernorm", "model.layers.32.mlp.gate.e_score_correction_bias", "model.layers.27.input_layernorm", "model.layers.16.post_attention_layernorm", "model.layers.35.self_attn.q_a_layernorm", "model.layers.58.self_attn.indexers_proj", "model.layers.3.mlp.gate.e_score_correction_bias", "model.layers.49.self_attn.q_a_layernorm", "model.layers.66.mlp.gate.e_score_correction_bias", "model.layers.10.self_attn.q_a_layernorm", "model.layers.51.mlp.gate.e_score_correction_bias", "model.layers.64.self_attn.q_a_layernorm", "model.layers.70.input_layernorm", "model.layers.71.post_attention_layernorm", "model.layers.67.post_attention_layernorm", "model.layers.6.self_attn.indexer.k_norm", "model.layers.3.mlp.gate", "model.layers.11.mlp.gate", "model.layers.17.input_layernorm", "model.layers.13.mlp.gate", "model.layers.15.mlp.gate", "model.layers.71.self_attn.q_a_layernorm", "model.layers.33.mlp.gate.e_score_correction_bias", "model.layers.72.mlp.gate.e_score_correction_bias", "model.layers.65.post_attention_layernorm", "model.layers.78.hnorm", "model.layers.40.post_attention_layernorm", "model.layers.12.mlp.gate.e_score_correction_bias", "model.layers.8.input_layernorm", "model.layers.73.mlp.gate", "model.layers.11.post_attention_layernorm", "model.layers.61.input_layernorm", "model.layers.14.self_attn.kv_a_layernorm", "model.layers.39.self_attn.kv_a_layernorm", "model.layers.70.self_attn.indexer.k_norm", "model.layers.45.mlp.gate", "model.layers.51.self_attn.q_a_layernorm", "model.layers.64.post_attention_layernorm", "model.layers.20.post_attention_layernorm", "model.layers.66.self_attn.kv_a_layernorm", "model.layers.10.self_attn.indexer.k_norm.bias", "model.layers.23.input_layernorm", "model.layers.38.self_attn.indexer.k_norm", "model.layers.19.self_attn.kv_a_layernorm", "model.layers.40.self_attn.q_a_layernorm", "model.layers.10.self_attn.indexer.k_norm", "model.layers.44.post_attention_layernorm", "model.layers.70.post_attention_layernorm", "model.layers.41.mlp.gate.e_score_correction_bias", "model.layers.36.post_attention_layernorm", "model.layers.30.self_attn.indexer.k_norm.bias", "model.layers.54.self_attn.q_a_layernorm", "model.layers.59.mlp.gate", "model.layers.69.post_attention_layernorm", "lm_head", "model.layers.77.self_attn.kv_a_layernorm", "model.layers.75.self_attn.q_a_layernorm", "model.layers.56.self_attn.q_a_layernorm", "model.layers.50.self_attn.indexer.k_norm.bias", "model.layers.54.mlp.gate", "model.layers.6.self_attn.indexer.k_norm.bias", "model.layers.15.input_layernorm", "model.layers.29.self_attn.q_a_layernorm", "model.layers.54.self_attn.indexers_proj", "model.layers.73.self_attn.kv_a_layernorm", "model.layers.0.self_attn.indexer.k_norm", "model.layers.58.self_attn.indexer.k_norm", "model.layers.36.input_layernorm", "model.layers.47.self_attn.q_a_layernorm", "model.layers.26.mlp.gate.e_score_correction_bias", "model.layers.50.input_layernorm", "model.layers.16.self_attn.q_a_layernorm", "model.layers.19.mlp.gate.e_score_correction_bias", "model.layers.78.input_layernorm", "model.layers.27.mlp.gate", "model.layers.19.self_attn.q_a_layernorm", "model.layers.25.input_layernorm", "model.layers.33.self_attn.kv_a_layernorm", "model.layers.32.mlp.gate", "model.layers.61.self_attn.kv_a_layernorm", "model.layers.50.self_attn.indexers_proj", "model.layers.34.self_attn.indexers_proj", "model.layers.4.mlp.gate.e_score_correction_bias", "model.layers.76.post_attention_layernorm", "model.layers.57.post_attention_layernorm", "model.layers.73.post_attention_layernorm", "model.layers.68.mlp.gate.e_score_correction_bias", "model.layers.34.self_attn.kv_a_layernorm", "model.layers.78.enorm", "model.layers.30.mlp.gate", "model.layers.10.self_attn.indexers_proj", "model.layers.34.input_layernorm", "model.layers.34.post_attention_layernorm", "model.layers.20.mlp.gate.e_score_correction_bias", "model.layers.54.self_attn.indexer.k_norm.bias", "model.layers.44.mlp.gate", "model.layers.25.mlp.gate.e_score_correction_bias", "model.layers.26.self_attn.indexers_proj", "model.layers.66.self_attn.indexer.k_norm.bias", "model.layers.41.post_attention_layernorm", "model.layers.51.post_attention_layernorm", "model.layers.22.self_attn.indexer.k_norm.bias", "model.layers.21.mlp.gate.e_score_correction_bias", "model.layers.15.self_attn.kv_a_layernorm", "model.layers.25.self_attn.kv_a_layernorm", "model.layers.55.mlp.gate", "model.layers.11.self_attn.q_a_layernorm", "model.layers.72.mlp.gate", "model.layers.21.self_attn.kv_a_layernorm", "model.layers.59.mlp.gate.e_score_correction_bias", "model.layers.2.input_layernorm", "model.layers.5.mlp.gate", "model.layers.54.self_attn.kv_a_layernorm", "model.layers.0.self_attn.indexers_proj", "model.layers.71.mlp.gate.e_score_correction_bias", "model.layers.6.self_attn.q_a_layernorm", "model.layers.14.mlp.gate", "model.layers.44.self_attn.kv_a_layernorm", "model.layers.61.post_attention_layernorm", "model.layers.64.self_attn.kv_a_layernorm", "model.layers.1.post_attention_layernorm", "model.layers.4.mlp.gate", "model.layers.22.post_attention_layernorm", "model.layers.21.self_attn.q_a_layernorm", "model.layers.74.self_attn.q_a_layernorm", "model.layers.18.self_attn.indexers_proj", "model.layers.75.input_layernorm", "model.layers.70.self_attn.q_a_layernorm", "model.layers.22.self_attn.indexer.k_norm", "model.layers.42.mlp.gate.e_score_correction_bias", "model.layers.5.self_attn.kv_a_layernorm", "model.layers.31.mlp.gate.e_score_correction_bias", "model.layers.62.self_attn.indexer.k_norm.bias", "model.layers.29.post_attention_layernorm", "model.layers.6.self_attn.kv_a_layernorm", "model.layers.48.input_layernorm", "model.layers.74.self_attn.indexer.k_norm.bias", "model.layers.7.input_layernorm", "model.layers.39.self_attn.q_a_layernorm", "model.layers.41.input_layernorm", "model.layers.54.self_attn.indexer.k_norm", "model.layers.32.self_attn.q_a_layernorm", "model.layers.39.mlp.gate.e_score_correction_bias", "model.layers.42.input_layernorm", "model.layers.43.mlp.gate", "model.layers.72.self_attn.q_a_layernorm", "model.layers.37.mlp.gate.e_score_correction_bias", "model.layers.5.post_attention_layernorm", "model.layers.28.mlp.gate", "model.layers.2.post_attention_layernorm", "model.layers.49.post_attention_layernorm", "model.layers.43.self_attn.kv_a_layernorm", "model.layers.12.input_layernorm", "model.layers.56.mlp.gate", "model.layers.5.input_layernorm", "model.layers.29.mlp.gate.e_score_correction_bias", "model.layers.4.self_attn.kv_a_layernorm", "model.layers.20.self_attn.q_a_layernorm", "model.layers.59.self_attn.q_a_layernorm", "model.layers.68.mlp.gate", "model.layers.21.post_attention_layernorm", "model.layers.26.self_attn.q_a_layernorm", "model.layers.52.self_attn.q_a_layernorm", "model.layers.29.self_attn.kv_a_layernorm", "model.layers.24.input_layernorm", "model.layers.34.self_attn.indexer.k_norm.bias", "model.layers.43.input_layernorm", "model.layers.27.self_attn.kv_a_layernorm", "model.layers.31.input_layernorm", "model.layers.69.mlp.gate", "model.layers.26.post_attention_layernorm", "model.layers.3.self_attn.q_a_layernorm", "model.layers.37.post_attention_layernorm", "model.layers.52.mlp.gate", "model.layers.73.input_layernorm", "model.layers.19.input_layernorm", "model.layers.13.self_attn.q_a_layernorm", "model.layers.76.input_layernorm", "model.layers.8.mlp.gate", "model.layers.63.post_attention_layernorm", "model.layers.58.mlp.gate", "model.layers.31.post_attention_layernorm", "model.layers.23.mlp.gate.e_score_correction_bias", "model.layers.34.self_attn.q_a_layernorm", "model.layers.78.self_attn.indexer.k_norm.bias", "model.layers.13.mlp.gate.e_score_correction_bias", "model.layers.71.mlp.gate", "model.layers.14.input_layernorm", "model.layers.17.mlp.gate.e_score_correction_bias", "model.layers.8.post_attention_layernorm", "model.layers.45.self_attn.kv_a_layernorm", "model.layers.46.self_attn.indexer.k_norm", "model.layers.20.self_attn.kv_a_layernorm", "model.layers.78.self_attn.indexer.k_norm", "model.layers.37.mlp.gate", "model.layers.7.mlp.gate", "model.layers.30.mlp.gate.e_score_correction_bias", "model.layers.67.mlp.gate", "model.layers.73.self_attn.q_a_layernorm", "model.layers.10.self_attn.kv_a_layernorm", "model.layers.61.mlp.gate", "model.layers.22.mlp.gate.e_score_correction_bias", "model.layers.58.input_layernorm", "model.layers.8.self_attn.kv_a_layernorm", "model.layers.33.self_attn.q_a_layernorm", "model.layers.55.self_attn.q_a_layernorm", "model.layers.32.input_layernorm", "model.layers.17.self_attn.kv_a_layernorm", "model.layers.46.self_attn.indexer.k_norm.bias", "model.layers.11.mlp.gate.e_score_correction_bias", "model.layers.9.input_layernorm", "model.layers.77.input_layernorm", "model.layers.16.self_attn.kv_a_layernorm", "model.layers.45.post_attention_layernorm", "model.layers.74.mlp.gate", "model.layers.68.self_attn.q_a_layernorm", "model.layers.55.self_attn.kv_a_layernorm", "model.layers.44.input_layernorm", "model.layers.60.self_attn.kv_a_layernorm", "model.layers.14.post_attention_layernorm", "model.layers.62.self_attn.indexers_proj", "model.layers.0.self_attn.q_a_layernorm", "model.layers.72.post_attention_layernorm", "model.layers.78.post_attention_layernorm", "model.layers.60.mlp.gate.e_score_correction_bias", "model.layers.26.self_attn.indexer.k_norm.bias", "model.layers.38.post_attention_layernorm", "model.layers.50.post_attention_layernorm", "model.layers.28.input_layernorm", "model.layers.30.self_attn.indexer.k_norm", "model.layers.37.self_attn.q_a_layernorm", "model.layers.56.input_layernorm", "model.layers.27.post_attention_layernorm", "model.layers.74.input_layernorm", "model.layers.78.self_attn.q_a_layernorm", "model.layers.42.self_attn.q_a_layernorm", "model.layers.26.self_attn.indexer.k_norm", "model.layers.42.self_attn.kv_a_layernorm", "model.layers.1.self_attn.indexer.k_norm", "model.layers.39.post_attention_layernorm", "model.layers.48.mlp.gate.e_score_correction_bias", "model.layers.74.post_attention_layernorm", "model.layers.69.self_attn.q_a_layernorm", "model.layers.13.input_layernorm", "model.layers.74.self_attn.indexer.k_norm", "model.layers.29.mlp.gate", "model.layers.7.mlp.gate.e_score_correction_bias", "model.layers.3.self_attn.kv_a_layernorm", "model.layers.35.self_attn.kv_a_layernorm", "model.layers.46.self_attn.indexers_proj", "model.layers.0.self_attn.kv_a_layernorm", "model.layers.43.self_attn.q_a_layernorm", "model.norm", "model.layers.1.self_attn.kv_a_layernorm", "model.layers.38.self_attn.indexer.k_norm.bias", "model.layers.30.self_attn.q_a_layernorm", "model.layers.0.input_layernorm", "model.layers.47.input_layernorm", "model.layers.8.self_attn.q_a_layernorm", "model.layers.61.self_attn.q_a_layernorm", "model.layers.1.self_attn.indexer.k_norm.bias", "model.layers.62.mlp.gate.e_score_correction_bias", "model.layers.66.self_attn.q_a_layernorm", "model.layers.46.self_attn.kv_a_layernorm", "model.layers.66.mlp.gate", "model.layers.69.input_layernorm", "model.layers.69.self_attn.kv_a_layernorm", "model.layers.22.input_layernorm", "model.layers.30.post_attention_layernorm", "model.layers.24.self_attn.q_a_layernorm", "model.layers.48.mlp.gate", "model.layers.7.self_attn.kv_a_layernorm", "model.layers.67.input_layernorm", "model.layers.27.self_attn.q_a_layernorm", "model.layers.23.mlp.gate", "model.layers.45.mlp.gate.e_score_correction_bias", "model.layers.76.self_attn.kv_a_layernorm", "model.layers.22.self_attn.q_a_layernorm", "model.layers.78.mlp.gate", "model.layers.9.mlp.gate", "model.layers.57.self_attn.q_a_layernorm", "model.layers.78.mlp.gate.e_score_correction_bias", "model.layers.23.post_attention_layernorm", "model.layers.38.self_attn.q_a_layernorm", "model.layers.53.mlp.gate", "model.layers.77.self_attn.q_a_layernorm", "model.layers.62.self_attn.kv_a_layernorm", "model.layers.12.mlp.gate", "model.layers.34.self_attn.indexer.k_norm", "model.layers.46.self_attn.q_a_layernorm", "model.layers.36.self_attn.q_a_layernorm", "model.layers.58.self_attn.indexer.k_norm.bias", "model.layers.60.post_attention_layernorm", "model.layers.28.post_attention_layernorm", "model.layers.2.self_attn.kv_a_layernorm", "model.layers.70.self_attn.indexers_proj", "model.layers.16.mlp.gate.e_score_correction_bias", "model.layers.2.self_attn.indexer.k_norm", "model.layers.31.self_attn.kv_a_layernorm", "model.layers.64.input_layernorm", "model.layers.6.mlp.gate.e_score_correction_bias", "model.layers.50.self_attn.indexer.k_norm", "model.layers.58.self_attn.q_a_layernorm", "model.layers.57.mlp.gate.e_score_correction_bias", "model.layers.13.post_attention_layernorm", "model.layers.28.self_attn.kv_a_layernorm", "model.layers.31.self_attn.q_a_layernorm", "model.layers.1.self_attn.q_a_layernorm", "model.layers.17.mlp.gate", "model.layers.15.post_attention_layernorm", "model.layers.18.self_attn.indexer.k_norm.bias", "model.layers.18.self_attn.indexer.k_norm", "model.layers.43.mlp.gate.e_score_correction_bias", "model.layers.11.input_layernorm", "model.layers.9.post_attention_layernorm", "model.layers.33.input_layernorm", "model.layers.66.input_layernorm", "model.layers.78.self_attn.indexers_proj", "model.layers.54.post_attention_layernorm", "model.layers.53.mlp.gate.e_score_correction_bias", "model.layers.5.mlp.gate.e_score_correction_bias", "model.layers.19.mlp.gate", "model.layers.49.mlp.gate", "model.layers.66.self_attn.indexers_proj", "model.layers.14.self_attn.indexer.k_norm.bias", "model.layers.41.mlp.gate", "model.layers.42.self_attn.indexer.k_norm.bias", "model.layers.65.input_layernorm", "model.layers.75.mlp.gate", "model.layers.12.self_attn.kv_a_layernorm", "model.layers.53.input_layernorm", "model.layers.58.self_attn.kv_a_layernorm", "model.layers.60.self_attn.q_a_layernorm", "model.layers.6.mlp.gate", "model.layers.49.self_attn.kv_a_layernorm", "model.layers.16.mlp.gate", "model.layers.22.mlp.gate", "model.layers.58.post_attention_layernorm", "model.layers.15.mlp.gate.e_score_correction_bias", "model.layers.26.input_layernorm", "model.layers.3.post_attention_layernorm", "model.layers.6.input_layernorm", "model.layers.22.self_attn.kv_a_layernorm", "model.layers.16.input_layernorm", "model.layers.33.post_attention_layernorm", "model.layers.17.post_attention_layernorm", "model.layers.35.mlp.gate", "model.layers.56.post_attention_layernorm", "model.layers.7.self_attn.q_a_layernorm", "model.layers.24.mlp.gate", "model.layers.24.post_attention_layernorm", "model.layers.36.mlp.gate.e_score_correction_bias", "model.layers.56.mlp.gate.e_score_correction_bias", "model.layers.32.post_attention_layernorm", "model.layers.47.post_attention_layernorm", "model.layers.40.mlp.gate", "model.layers.41.self_attn.kv_a_layernorm", "model.layers.52.self_attn.kv_a_layernorm", "model.layers.9.self_attn.kv_a_layernorm", "model.layers.77.mlp.gate.e_score_correction_bias", "model.layers.78.shared_head.norm", "model.layers.53.post_attention_layernorm", "model.layers.63.self_attn.kv_a_layernorm", "model.layers.48.self_attn.kv_a_layernorm", "model.layers.28.mlp.gate.e_score_correction_bias", "model.layers.8.mlp.gate.e_score_correction_bias", "model.layers.52.mlp.gate.e_score_correction_bias", "model.layers.52.post_attention_layernorm", "model.layers.18.self_attn.q_a_layernorm", "model.layers.50.mlp.gate.e_score_correction_bias", "model.layers.64.mlp.gate.e_score_correction_bias", "model.layers.38.mlp.gate.e_score_correction_bias", "model.layers.59.input_layernorm", "model.layers.21.input_layernorm", "model.layers.39.mlp.gate", "model.layers.77.post_attention_layernorm", "model.layers.76.self_attn.q_a_layernorm", "model.layers.12.post_attention_layernorm", "model.layers.43.post_attention_layernorm", "model.layers.35.input_layernorm", "model.layers.62.mlp.gate", "model.layers.38.self_attn.kv_a_layernorm", "model.layers.63.mlp.gate", "model.layers.27.mlp.gate.e_score_correction_bias", "model.layers.18.mlp.gate", "model.layers.25.post_attention_layernorm", "model.layers.55.mlp.gate.e_score_correction_bias", "model.layers.66.self_attn.indexer.k_norm", "model.layers.48.self_attn.q_a_layernorm", "model.layers.78.eh_proj", "model.layers.77.mlp.gate", "model.layers.6.self_attn.indexers_proj", "model.layers.14.self_attn.q_a_layernorm", "model.layers.2.self_attn.q_a_layernorm", "model.layers.42.mlp.gate", "model.layers.54.mlp.gate.e_score_correction_bias", "model.layers.60.input_layernorm", "model.layers.40.self_attn.kv_a_layernorm", "model.layers.51.self_attn.kv_a_layernorm", "model.layers.18.input_layernorm", "model.layers.30.self_attn.indexers_proj", "model.layers.53.self_attn.kv_a_layernorm", "model.layers.6.post_attention_layernorm", "model.layers.70.self_attn.indexer.k_norm.bias", "model.layers.54.input_layernorm", "model.layers.41.self_attn.q_a_layernorm", "model.layers.18.mlp.gate.e_score_correction_bias", "model.layers.34.mlp.gate.e_score_correction_bias", "model.layers.47.mlp.gate", "model.layers.46.mlp.gate", "model.layers.25.self_attn.q_a_layernorm", "model.layers.63.mlp.gate.e_score_correction_bias", "model.layers.66.post_attention_layernorm", "model.layers.76.mlp.gate", "model.embed_tokens", "model.layers.67.self_attn.kv_a_layernorm", "model.layers.67.mlp.gate.e_score_correction_bias", "model.layers.26.self_attn.kv_a_layernorm", "model.layers.31.mlp.gate", "model.layers.18.self_attn.kv_a_layernorm", "model.layers.34.mlp.gate", "model.layers.55.post_attention_layernorm", "model.layers.30.input_layernorm", "model.layers.50.mlp.gate", "model.layers.63.input_layernorm", "model.layers.11.self_attn.kv_a_layernorm", "model.layers.24.self_attn.kv_a_layernorm", "model.layers.75.mlp.gate.e_score_correction_bias", "model.layers.10.post_attention_layernorm", "model.layers.1.self_attn.indexers_proj", "model.layers.42.self_attn.indexers_proj", "model.layers.46.input_layernorm", "model.layers.23.self_attn.q_a_layernorm", "model.layers.63.self_attn.q_a_layernorm", "model.layers.14.self_attn.indexer.k_norm", "model.layers.14.self_attn.indexers_proj", "model.layers.60.mlp.gate", "model.layers.70.mlp.gate.e_score_correction_bias", "model.layers.53.self_attn.q_a_layernorm", "model.layers.0.post_attention_layernorm", "model.layers.35.mlp.gate.e_score_correction_bias", "model.layers.75.self_attn.kv_a_layernorm", "model.layers.1.input_layernorm", "model.layers.42.post_attention_layernorm", "model.layers.37.input_layernorm", "model.layers.9.mlp.gate.e_score_correction_bias", "model.layers.46.post_attention_layernorm", "model.layers.26.mlp.gate", "model.layers.65.mlp.gate.e_score_correction_bias", "model.layers.67.self_attn.q_a_layernorm", "model.layers.44.mlp.gate.e_score_correction_bias", "model.layers.33.mlp.gate", "model.layers.29.input_layernorm", "model.layers.18.post_attention_layernorm", "model.layers.75.post_attention_layernorm", "model.layers.44.self_attn.q_a_layernorm", "model.layers.10.mlp.gate", "model.layers.21.mlp.gate", "model.layers.58.mlp.gate.e_score_correction_bias", "model.layers.7.post_attention_layernorm", "model.layers.50.self_attn.q_a_layernorm", "model.layers.9.self_attn.q_a_layernorm", "model.layers.2.self_attn.indexers_proj", "model.layers.0.self_attn.indexer.k_norm.bias", "model.layers.74.self_attn.kv_a_layernorm", "model.layers.74.mlp.gate.e_score_correction_bias", "model.layers.49.mlp.gate.e_score_correction_bias", "model.layers.62.post_attention_layernorm", "model.layers.57.self_attn.kv_a_layernorm", "model.layers.40.input_layernorm", "model.layers.69.mlp.gate.e_score_correction_bias", "model.layers.35.post_attention_layernorm", "model.layers.74.self_attn.indexers_proj", "model.layers.55.input_layernorm", "model.layers.49.input_layernorm", "model.layers.40.mlp.gate.e_score_correction_bias", "model.layers.25.mlp.gate", "model.layers.2.self_attn.indexer.k_norm.bias", "model.layers.68.self_attn.kv_a_layernorm", "model.layers.51.input_layernorm", "model.layers.22.self_attn.indexers_proj", "model.layers.10.input_layernorm", "model.layers.42.self_attn.indexer.k_norm", "model.layers.56.self_attn.kv_a_layernorm", "model.layers.17.self_attn.q_a_layernorm", "model.layers.59.self_attn.kv_a_layernorm", "model.layers.3.input_layernorm", "model.layers.57.input_layernorm", "model.layers.38.input_layernorm", "model.layers.62.self_attn.indexer.k_norm", "model.layers.20.mlp.gate", "model.layers.64.mlp.gate", "model.layers.28.self_attn.q_a_layernorm", "model.layers.32.self_attn.kv_a_layernorm", "model.layers.10.mlp.gate.e_score_correction_bias", "model.layers.5.self_attn.q_a_layernorm", "model.layers.37.self_attn.kv_a_layernorm" ], "indexMetadataTotalSize": 755617140416, "indexTensorCount": 118629, "metadataSafetensors": { "parameters": { "F32": 45872560, "BF16": 2103729152, "F8_E4M3": 751226191872 }, "total": 753375793584 } }, "checkpointAudit": { "statement": "Each listed result is derived from an immutable official checkpoint tensor read by exact HTTP byte range. Raw tensor values are discarded; only hashes, byte ranges, and aggregates are archived.", "tensors": [ { "layer": 3, "tensor": "model.layers.3.mlp.gate.e_score_correction_bias", "shard": "model-00040-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105136, "headerSha256": "c49829bf9993a98c0c22633000bf1d8092f4f3e4c5d2f7c907128e1ae8c042d8", "tensorByteRange": [ 836280, 837303 ], "tensorBytes": 1024, "tensorSha256": "eb6feeb8d7ab446e4e786aaac55c22cc7b98521dbd71cb0a57610d8da59b0491", "rangeHttpStatus": 206, "contentRange": "bytes 836280-837303/5366353016", "valueCount": 256, "min": 33.985321044921875, "max": 34.62297058105469, "spread": 0.6376495361328125, "mean": 34.45753416419029, "standardDeviation": 0.13826611249206686, "fp32Distinct": 174, "bf16Distinct": 3, "bf16Min": 34, "bf16Max": 34.5, "meanAbsoluteQuantizationError": 0.06975498795509338, "maxAbsoluteQuantizationError": 0.1248931884765625, "zeroLogitTop8SetDifference": 8 }, { "layer": 4, "tensor": "model.layers.4.mlp.gate.e_score_correction_bias", "shard": "model-00060-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 104896, "headerSha256": "843648c3c4fd6c90bce4aa6aef3fb9cd4667baa24df304beb31a22a46a1b1d88", "tensorByteRange": [ 1167816, 1168839 ], "tensorBytes": 1024, "tensorSha256": "0e0c15e8768c258bac3334e4cc7b7cb6c1ece46f0af6cbee69485b3aa287678d", "rangeHttpStatus": 206, "contentRange": "bytes 1167816-1168839/5366352776", "valueCount": 256, "min": 28.102554321289062, "max": 28.66924285888672, "spread": 0.5666885375976562, "mean": 28.538271874189377, "standardDeviation": 0.14624599650953196, "fp32Distinct": 157, "bf16Distinct": 5, "bf16Min": 28.125, "bf16Max": 28.625, "meanAbsoluteQuantizationError": 0.022958189249038696, "maxAbsoluteQuantizationError": 0.06169891357421875, "zeroLogitTop8SetDifference": 8 }, { "layer": 5, "tensor": "model.layers.5.mlp.gate.e_score_correction_bias", "shard": "model-00081-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105992, "headerSha256": "a1c217ff45611ead4eed03aa8f7f7e790be7a3e2781a76044e9b6610508963e4", "tensorByteRange": [ 185872, 186895 ], "tensorBytes": 1024, "tensorSha256": "966371807bfd4493bc5aada7453635d0d0fab7c771c80a137630e217181ad5c8", "rangeHttpStatus": 206, "contentRange": "bytes 185872-186895/5366353872", "valueCount": 256, "min": 21.3162841796875, "max": 21.813011169433594, "spread": 0.49672698974609375, "mean": 21.69507598876953, "standardDeviation": 0.11296977748153715, "fp32Distinct": 152, "bf16Distinct": 5, "bf16Min": 21.375, "bf16Max": 21.875, "meanAbsoluteQuantizationError": 0.026111304759979248, "maxAbsoluteQuantizationError": 0.06198883056640625, "zeroLogitTop8SetDifference": 7 }, { "layer": 6, "tensor": "model.layers.6.mlp.gate.e_score_correction_bias", "shard": "model-00101-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106128, "headerSha256": "e5bb71099bc92c8637aca785f31f0ed1bfec4bd7bed464ec1c63bc6b602c4c67", "tensorByteRange": [ 514712, 515735 ], "tensorBytes": 1024, "tensorSha256": "088f6cbccdca7961727c51c557f7291e3d9fa30492d6b8a02f3d31f332ba2339", "rangeHttpStatus": 206, "contentRange": "bytes 514712-515735/5363339032", "valueCount": 256, "min": 15.210033416748047, "max": 15.686777114868164, "spread": 0.4767436981201172, "mean": 15.58792944252491, "standardDeviation": 0.086913244787114, "fp32Distinct": 229, "bf16Distinct": 9, "bf16Min": 15.1875, "bf16Max": 15.6875, "meanAbsoluteQuantizationError": 0.016736477613449097, "maxAbsoluteQuantizationError": 0.031164169311523438, "zeroLogitTop8SetDifference": 6 }, { "layer": 7, "tensor": "model.layers.7.mlp.gate.e_score_correction_bias", "shard": "model-00121-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105136, "headerSha256": "49f7f1b6aaf9ee4849705ab64c41a036dd1a56cea823c877b026f9faddc83d0b", "tensorByteRange": [ 842424, 843447 ], "tensorBytes": 1024, "tensorSha256": "5bc38af32823c75a2970e6787f116d71122798d2cf128ec3d76537d9ed4c892d", "rangeHttpStatus": 206, "contentRange": "bytes 842424-843447/5366353016", "valueCount": 256, "min": 11.413963317871094, "max": 11.670812606811523, "spread": 0.2568492889404297, "mean": 11.58847464621067, "standardDeviation": 0.05464490233269021, "fp32Distinct": 241, "bf16Distinct": 5, "bf16Min": 11.4375, "bf16Max": 11.6875, "meanAbsoluteQuantizationError": 0.01676110178232193, "maxAbsoluteQuantizationError": 0.031145095825195312, "zeroLogitTop8SetDifference": 6 }, { "layer": 8, "tensor": "model.layers.8.mlp.gate.e_score_correction_bias", "shard": "model-00140-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105280, "headerSha256": "d7fd81466c9034639430e0b73b9045d3a7c5110a27b8f94327c2d728a7356b98", "tensorByteRange": [ 108360, 109383 ], "tensorBytes": 1024, "tensorSha256": "93f405a383dc621f6231e152caccb8f42c1c6c750845272aa09bbd55401bb4f6", "rangeHttpStatus": 206, "contentRange": "bytes 108360-109383/5366353160", "valueCount": 256, "min": 9.467058181762695, "max": 9.85483169555664, "spread": 0.3877735137939453, "mean": 9.732149846851826, "standardDeviation": 0.04698851551130864, "fp32Distinct": 235, "bf16Distinct": 7, "bf16Min": 9.4375, "bf16Max": 9.875, "meanAbsoluteQuantizationError": 0.015261299908161163, "maxAbsoluteQuantizationError": 0.031097412109375, "zeroLogitTop8SetDifference": 5 }, { "layer": 9, "tensor": "model.layers.9.mlp.gate.e_score_correction_bias", "shard": "model-00141-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 91896, "headerSha256": "c365cdc0d959fdd81e83a7dfedb3eb21ee8eb751d138a868854049eb0a941465", "tensorByteRange": [ 1194752, 1195775 ], "tensorBytes": 1024, "tensorSha256": "55c9c9adbe16d37fc948a74be492082b3d30ee4e4cf483fffd26604a19fec452", "rangeHttpStatus": 206, "contentRange": "bytes 1194752-1195775/4724454592", "valueCount": 256, "min": 6.826382637023926, "max": 7.105213165283203, "spread": 0.27883052825927734, "mean": 7.019456852227449, "standardDeviation": 0.040455044068427114, "fp32Distinct": 223, "bf16Distinct": 10, "bf16Min": 6.8125, "bf16Max": 7.09375, "meanAbsoluteQuantizationError": 0.007378779351711273, "maxAbsoluteQuantizationError": 0.015535354614257812, "zeroLogitTop8SetDifference": 4 }, { "layer": 10, "tensor": "model.layers.10.mlp.gate.e_score_correction_bias", "shard": "model-00003-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106216, "headerSha256": "2d19bdad64ed13325c2478e55ad4e5c054ae6b94e91e56cdac1428b12b77bb65", "tensorByteRange": [ 972528, 973551 ], "tensorBytes": 1024, "tensorSha256": "907a7863bd51c582a81dea96318f5bebf4cd87cff1239d8e43d489705bf08804", "rangeHttpStatus": 206, "contentRange": "bytes 972528-973551/5363339120", "valueCount": 256, "min": 6.817388534545898, "max": 7.063243865966797, "spread": 0.24585533142089844, "mean": 6.9855689108371735, "standardDeviation": 0.0402298797751221, "fp32Distinct": 238, "bf16Distinct": 8, "bf16Min": 6.8125, "bf16Max": 7.0625, "meanAbsoluteQuantizationError": 0.007821537554264069, "maxAbsoluteQuantizationError": 0.015540122985839844, "zeroLogitTop8SetDifference": 4 }, { "layer": 11, "tensor": "model.layers.11.mlp.gate.e_score_correction_bias", "shard": "model-00005-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105680, "headerSha256": "5c35b24dac21e2ff4106d4b4a44685c4703d4c19f76e637fc13d94a2e56e24a4", "tensorByteRange": [ 766168, 767191 ], "tensorBytes": 1024, "tensorSha256": "a6d32d06907d2a8f2f5de273f8eb833712a165124e0866e52637410b0205e2b6", "rangeHttpStatus": 206, "contentRange": "bytes 766168-767191/5366353560", "valueCount": 256, "min": 6.804360389709473, "max": 7.019268989562988, "spread": 0.21490859985351562, "mean": 6.942926308140159, "standardDeviation": 0.034248940100057745, "fp32Distinct": 235, "bf16Distinct": 8, "bf16Min": 6.8125, "bf16Max": 7.03125, "meanAbsoluteQuantizationError": 0.00803522951900959, "maxAbsoluteQuantizationError": 0.0156097412109375, "zeroLogitTop8SetDifference": 4 }, { "layer": 12, "tensor": "model.layers.12.mlp.gate.e_score_correction_bias", "shard": "model-00007-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105824, "headerSha256": "91198156b65e402bbd23ab9750042621c3ab24dcedc0240d4a27abfa01f2b56c", "tensorByteRange": [ 557416, 558439 ], "tensorBytes": 1024, "tensorSha256": "1620597d70bfe213342a10d1238991127fa9580ea3ea31bc9218270a74884e54", "rangeHttpStatus": 206, "contentRange": "bytes 557416-558439/5366353704", "valueCount": 256, "min": 7.8098649978637695, "max": 8.10773754119873, "spread": 0.29787254333496094, "mean": 8.037340292707086, "standardDeviation": 0.03787155279806369, "fp32Distinct": 221, "bf16Distinct": 8, "bf16Min": 7.8125, "bf16Max": 8.125, "meanAbsoluteQuantizationError": 0.015197178348898888, "maxAbsoluteQuantizationError": 0.030759811401367188, "zeroLogitTop8SetDifference": 6 }, { "layer": 13, "tensor": "model.layers.13.mlp.gate.e_score_correction_bias", "shard": "model-00009-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105952, "headerSha256": "a1cf2ae27d09ddc4340589addc9696dbb522cac23f3341f4898e3b76dce36a51", "tensorByteRange": [ 348648, 349671 ], "tensorBytes": 1024, "tensorSha256": "c9a735594b1476ec3444bfb84461f65ee19bf62f048cc2378fe1b32a1185ef2b", "rangeHttpStatus": 206, "contentRange": "bytes 348648-349671/5366353832", "valueCount": 256, "min": 6.018793106079102, "max": 6.24169921875, "spread": 0.22290611267089844, "mean": 6.203384837135673, "standardDeviation": 0.024296777942044698, "fp32Distinct": 196, "bf16Distinct": 7, "bf16Min": 6.03125, "bf16Max": 6.25, "meanAbsoluteQuantizationError": 0.0066861677914857864, "maxAbsoluteQuantizationError": 0.015489578247070312, "zeroLogitTop8SetDifference": 5 }, { "layer": 14, "tensor": "model.layers.14.mlp.gate.e_score_correction_bias", "shard": "model-00011-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106696, "headerSha256": "7cf0fda5555d703b0315622d9f5a0cb53a4235f981961b06c0b03f390a09c1c9", "tensorByteRange": [ 140496, 141519 ], "tensorBytes": 1024, "tensorSha256": "4b33d11373f9a4040270775a30c132d40add0f3a3d98fe15a56cd373aa07fb32", "rangeHttpStatus": 206, "contentRange": "bytes 140496-141519/5363339600", "valueCount": 256, "min": 5.907900810241699, "max": 6.109773635864258, "spread": 0.2018728256225586, "mean": 6.073685400187969, "standardDeviation": 0.02700875436227182, "fp32Distinct": 201, "bf16Distinct": 7, "bf16Min": 5.90625, "bf16Max": 6.125, "meanAbsoluteQuantizationError": 0.007132992148399353, "maxAbsoluteQuantizationError": 0.015551090240478516, "zeroLogitTop8SetDifference": 6 }, { "layer": 15, "tensor": "model.layers.15.mlp.gate.e_score_correction_bias", "shard": "model-00012-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105608, "headerSha256": "324f7dad5ba8bcfd39783e7a1adab9029375feff122b1bd60c730b6f99117560", "tensorByteRange": [ 1242256, 1243279 ], "tensorBytes": 1024, "tensorSha256": "41a81ec43b9dac15f3f790cdec6575ba0918f040dd71e78652bd8038e4e7a71e", "rangeHttpStatus": 206, "contentRange": "bytes 1242256-1243279/5366353488", "valueCount": 256, "min": 7.9377288818359375, "max": 8.198747634887695, "spread": 0.2610187530517578, "mean": 8.149367310106754, "standardDeviation": 0.029864213046491655, "fp32Distinct": 220, "bf16Distinct": 5, "bf16Min": 7.9375, "bf16Max": 8.1875, "meanAbsoluteQuantizationError": 0.018213465809822083, "maxAbsoluteQuantizationError": 0.030767440795898438, "zeroLogitTop8SetDifference": 8 }, { "layer": 16, "tensor": "model.layers.16.mlp.gate.e_score_correction_bias", "shard": "model-00014-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105616, "headerSha256": "85e687addd5c6d3a5d8c783d3304e7dc1abd1ccb87e428bfd713fe4c2bccaa6e", "tensorByteRange": [ 1033368, 1034391 ], "tensorBytes": 1024, "tensorSha256": "5d921a8bf37032fb4a5bb517d48c9845ee7f6688c06aca25ed68b7552cd66d6a", "rangeHttpStatus": 206, "contentRange": "bytes 1033368-1034391/5366353496", "valueCount": 256, "min": 6.787381172180176, "max": 6.999295234680176, "spread": 0.2119140625, "mean": 6.952074244618416, "standardDeviation": 0.02782523083106838, "fp32Distinct": 216, "bf16Distinct": 7, "bf16Min": 6.78125, "bf16Max": 7, "meanAbsoluteQuantizationError": 0.008014518767595291, "maxAbsoluteQuantizationError": 0.015551567077636719, "zeroLogitTop8SetDifference": 0 }, { "layer": 17, "tensor": "model.layers.17.mlp.gate.e_score_correction_bias", "shard": "model-00016-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105640, "headerSha256": "9f408385a30bbcfb3897cd5b4ffd5e94952436207d5430182c7cfc880b245cd4", "tensorByteRange": [ 824496, 825519 ], "tensorBytes": 1024, "tensorSha256": "f738cda63a9c9fc1ddc1e6eb612694644c84875900c70e3d015fff7973a4cb45", "rangeHttpStatus": 206, "contentRange": "bytes 824496-825519/5366353520", "valueCount": 256, "min": 7.560953140258789, "max": 7.745889663696289, "spread": 0.1849365234375, "mean": 7.704163217917085, "standardDeviation": 0.02637467353922727, "fp32Distinct": 207, "bf16Distinct": 7, "bf16Min": 7.5625, "bf16Max": 7.75, "meanAbsoluteQuantizationError": 0.007401956245303154, "maxAbsoluteQuantizationError": 0.015581130981445312, "zeroLogitTop8SetDifference": 4 }, { "layer": 18, "tensor": "model.layers.18.mlp.gate.e_score_correction_bias", "shard": "model-00018-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106360, "headerSha256": "47861cc6836776123a9ee5bb09ae2ac304301825cd48dffdee98af4810d4736a", "tensorByteRange": [ 616320, 617343 ], "tensorBytes": 1024, "tensorSha256": "0e03da1ff1879a5c171e5c986e95210371f2d0914dd6329b402d3160aa563f47", "rangeHttpStatus": 206, "contentRange": "bytes 616320-617343/5363339264", "valueCount": 256, "min": 7.098214149475098, "max": 7.268155097961426, "spread": 0.16994094848632812, "mean": 7.235249264165759, "standardDeviation": 0.023269740968893615, "fp32Distinct": 202, "bf16Distinct": 6, "bf16Min": 7.09375, "bf16Max": 7.28125, "meanAbsoluteQuantizationError": 0.006707390770316124, "maxAbsoluteQuantizationError": 0.015570640563964844, "zeroLogitTop8SetDifference": 6 }, { "layer": 19, "tensor": "model.layers.19.mlp.gate.e_score_correction_bias", "shard": "model-00020-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 101000, "headerSha256": "6b83c5f9a368b626e09402f63130c9bfce587f88bb4088b46c836f9e46d8409a", "tensorByteRange": [ 405136, 406159 ], "tensorBytes": 1024, "tensorSha256": "1c24a8e7d670236a4e9a7c76e96e344d78fcfd837b5beb96b779a93c0d0863ec", "rangeHttpStatus": 206, "contentRange": "bytes 405136-406159/5364805840", "valueCount": 256, "min": 7.000240325927734, "max": 7.220180511474609, "spread": 0.219940185546875, "mean": 7.188002470880747, "standardDeviation": 0.02714126659623686, "fp32Distinct": 198, "bf16Distinct": 8, "bf16Min": 7, "bf16Max": 7.21875, "meanAbsoluteQuantizationError": 0.008923914283514023, "maxAbsoluteQuantizationError": 0.015562057495117188, "zeroLogitTop8SetDifference": 7 }, { "layer": 20, "tensor": "model.layers.20.mlp.gate.e_score_correction_bias", "shard": "model-00022-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105992, "headerSha256": "ba2a69e9cd8883b92f5c753592afd9caf9947f1d4574cf740e39c94f8d34fafa", "tensorByteRange": [ 299536, 300559 ], "tensorBytes": 1024, "tensorSha256": "2291bbcb9f26cb0a90f1b63dac72a3720efa32e30a860c3f2e65e6f881eb0753", "rangeHttpStatus": 206, "contentRange": "bytes 299536-300559/5366353872", "valueCount": 256, "min": 9.078277587890625, "max": 9.275209426879883, "spread": 0.1969318389892578, "mean": 9.245602067559958, "standardDeviation": 0.026817238310764546, "fp32Distinct": 208, "bf16Distinct": 4, "bf16Min": 9.0625, "bf16Max": 9.25, "meanAbsoluteQuantizationError": 0.011887285858392715, "maxAbsoluteQuantizationError": 0.03070831298828125, "zeroLogitTop8SetDifference": 8 }, { "layer": 21, "tensor": "model.layers.21.mlp.gate.e_score_correction_bias", "shard": "model-00023-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 107488, "headerSha256": "a36576fbd4db4d84e8029d6b6b174b122035db23cc58b92c67494f34f4cc0301", "tensorByteRange": [ 1400808, 1401831 ], "tensorBytes": 1024, "tensorSha256": "96a0486738c954f9aa4b06d98514555b8a7121f7c2d7c9969d168e9c0d2d1b41", "rangeHttpStatus": 206, "contentRange": "bytes 1400808-1401831/5357948328", "valueCount": 256, "min": 7.241132736206055, "max": 7.3900957107543945, "spread": 0.14896297454833984, "mean": 7.363955290988088, "standardDeviation": 0.02203456774572723, "fp32Distinct": 199, "bf16Distinct": 5, "bf16Min": 7.25, "bf16Max": 7.375, "meanAbsoluteQuantizationError": 0.006770173087716103, "maxAbsoluteQuantizationError": 0.015372276306152344, "zeroLogitTop8SetDifference": 6 }, { "layer": 22, "tensor": "model.layers.22.mlp.gate.e_score_correction_bias", "shard": "model-00025-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106200, "headerSha256": "318be4c621196a096db727a22b5188179fe1840f5c2f765d0d3546436d2d3b87", "tensorByteRange": [ 1193696, 1194719 ], "tensorBytes": 1024, "tensorSha256": "6074758bcf144c4110d8bdf2566a40b5839f71509748164a37234018288cf584", "rangeHttpStatus": 206, "contentRange": "bytes 1193696-1194719/5363339104", "valueCount": 256, "min": 6.461555480957031, "max": 6.635494232177734, "spread": 0.17393875122070312, "mean": 6.6073816902935505, "standardDeviation": 0.023431503992552637, "fp32Distinct": 203, "bf16Distinct": 5, "bf16Min": 6.46875, "bf16Max": 6.625, "meanAbsoluteQuantizationError": 0.007563058286905289, "maxAbsoluteQuantizationError": 0.015496253967285156, "zeroLogitTop8SetDifference": 8 }, { "layer": 23, "tensor": "model.layers.23.mlp.gate.e_score_correction_bias", "shard": "model-00027-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105624, "headerSha256": "4a21d994b267bd37cc62aabc49a4d6113f3fb8825eb23d0e9666438435c92efa", "tensorByteRange": [ 987296, 988319 ], "tensorBytes": 1024, "tensorSha256": "a84a9d3d2139d3d524ff6321fea4fb46291e307c55fef4999e62d22562080b32", "rangeHttpStatus": 206, "contentRange": "bytes 987296-988319/5366353504", "valueCount": 256, "min": 4.731479644775391, "max": 4.978403091430664, "spread": 0.24692344665527344, "mean": 4.950802402570844, "standardDeviation": 0.02767855206345575, "fp32Distinct": 191, "bf16Distinct": 6, "bf16Min": 4.71875, "bf16Max": 4.96875, "meanAbsoluteQuantizationError": 0.006515735760331154, "maxAbsoluteQuantizationError": 0.015336990356445312, "zeroLogitTop8SetDifference": 8 }, { "layer": 24, "tensor": "model.layers.24.mlp.gate.e_score_correction_bias", "shard": "model-00029-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105680, "headerSha256": "d4b47e120cc1d7c6431f0a4432241013254f1ee31309d0aa6ba591e0d6420677", "tensorByteRange": [ 778456, 779479 ], "tensorBytes": 1024, "tensorSha256": "3687a8df0c1c1ac4f26388ce3ba75747dc498eb270ea73b3ad8740d1d36d9a58", "rangeHttpStatus": 206, "contentRange": "bytes 778456-779479/5366353560", "valueCount": 256, "min": 5.318190574645996, "max": 5.51511287689209, "spread": 0.19692230224609375, "mean": 5.484764544293284, "standardDeviation": 0.028639698307277223, "fp32Distinct": 201, "bf16Distinct": 7, "bf16Min": 5.3125, "bf16Max": 5.5, "meanAbsoluteQuantizationError": 0.007001454010605812, "maxAbsoluteQuantizationError": 0.015610694885253906, "zeroLogitTop8SetDifference": 8 }, { "layer": 25, "tensor": "model.layers.25.mlp.gate.e_score_correction_bias", "shard": "model-00031-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105808, "headerSha256": "64af808a9b3c2f6c5299722ee16c82c1f0d7f3a0013fd3b5b7715c66508d076f", "tensorByteRange": [ 569688, 570711 ], "tensorBytes": 1024, "tensorSha256": "49c3981783955fe5e1bbff69db991db5cc5b7726a25d64f2f08de856598bc6ae", "rangeHttpStatus": 206, "contentRange": "bytes 569688-570711/5366353688", "valueCount": 256, "min": 5.656939506530762, "max": 5.941862106323242, "spread": 0.28492259979248047, "mean": 5.9081093948334455, "standardDeviation": 0.03230829933925685, "fp32Distinct": 214, "bf16Distinct": 7, "bf16Min": 5.65625, "bf16Max": 5.9375, "meanAbsoluteQuantizationError": 0.0069683995097875595, "maxAbsoluteQuantizationError": 0.015625, "zeroLogitTop8SetDifference": 7 }, { "layer": 26, "tensor": "model.layers.26.mlp.gate.e_score_correction_bias", "shard": "model-00033-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106536, "headerSha256": "bfdcc08a2cc58130d0221004f7f7308391ec6af8f0ff76351ce717c15aed72fb", "tensorByteRange": [ 361520, 362543 ], "tensorBytes": 1024, "tensorSha256": "d662fc133faca73e2fed2b2cd73f88d135d5846ba50a1de5fe3f7be063a9af7d", "rangeHttpStatus": 206, "contentRange": "bytes 361520-362543/5363339440", "valueCount": 256, "min": 4.660510063171387, "max": 4.838448524475098, "spread": 0.17793846130371094, "mean": 4.799048218876123, "standardDeviation": 0.029676028751915627, "fp32Distinct": 221, "bf16Distinct": 6, "bf16Min": 4.65625, "bf16Max": 4.84375, "meanAbsoluteQuantizationError": 0.008527662605047226, "maxAbsoluteQuantizationError": 0.015490531921386719, "zeroLogitTop8SetDifference": 5 }, { "layer": 27, "tensor": "model.layers.27.mlp.gate.e_score_correction_bias", "shard": "model-00035-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106096, "headerSha256": "1970458b804e62f2a7f132d0efe2dc41e027188465d914d31eda1c7db64c56dd", "tensorByteRange": [ 155256, 156279 ], "tensorBytes": 1024, "tensorSha256": "6465fba852fa065aa31796f5910d53417d015bdd413c4cc5049d811410e318d6", "rangeHttpStatus": 206, "contentRange": "bytes 155256-156279/5366353976", "valueCount": 256, "min": 5.421137809753418, "max": 5.639023780822754, "spread": 0.21788597106933594, "mean": 5.599331920966506, "standardDeviation": 0.03776566915652802, "fp32Distinct": 224, "bf16Distinct": 8, "bf16Min": 5.40625, "bf16Max": 5.625, "meanAbsoluteQuantizationError": 0.007341621443629265, "maxAbsoluteQuantizationError": 0.015545845031738281, "zeroLogitTop8SetDifference": 8 }, { "layer": 28, "tensor": "model.layers.28.mlp.gate.e_score_correction_bias", "shard": "model-00036-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105608, "headerSha256": "2596170f8d34add1cde39353ef9353091409635af3d5a62eeec86e3cd3dcf45c", "tensorByteRange": [ 1254544, 1255567 ], "tensorBytes": 1024, "tensorSha256": "87d6bdb0cd2e2c823356b053e8bd187fd5273e956de86f960507c9c09025c25a", "rangeHttpStatus": 206, "contentRange": "bytes 1254544-1255567/5366353488", "valueCount": 256, "min": 4.34566593170166, "max": 4.592588424682617, "spread": 0.24692249298095703, "mean": 4.551389675587416, "standardDeviation": 0.03601920536324922, "fp32Distinct": 226, "bf16Distinct": 8, "bf16Min": 4.34375, "bf16Max": 4.59375, "meanAbsoluteQuantizationError": 0.00827062875032425, "maxAbsoluteQuantizationError": 0.015622138977050781, "zeroLogitTop8SetDifference": 6 }, { "layer": 29, "tensor": "model.layers.29.mlp.gate.e_score_correction_bias", "shard": "model-00038-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105424, "headerSha256": "8607b1b281f5ca8c7b166376e8f6d7eb9ca07f79200f6095f0f55ca35149ba56", "tensorByteRange": [ 1045464, 1046487 ], "tensorBytes": 1024, "tensorSha256": "fc7627d0e52a7ba578d7c9b0fd53c5f24505ce15a3647a10f63135b7b786bd62", "rangeHttpStatus": 206, "contentRange": "bytes 1045464-1046487/5366353304", "valueCount": 256, "min": 5.3051300048828125, "max": 5.6380157470703125, "spread": 0.3328857421875, "mean": 5.5810123439878225, "standardDeviation": 0.04696252395676736, "fp32Distinct": 232, "bf16Distinct": 9, "bf16Min": 5.3125, "bf16Max": 5.625, "meanAbsoluteQuantizationError": 0.007875548675656319, "maxAbsoluteQuantizationError": 0.015550613403320312, "zeroLogitTop8SetDifference": 7 }, { "layer": 30, "tensor": "model.layers.30.mlp.gate.e_score_correction_bias", "shard": "model-00042-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106360, "headerSha256": "b1117af6aab8d7bb6ae23b13f45021e48773464ac33684b964f6ecb8c7f39997", "tensorByteRange": [ 628608, 629631 ], "tensorBytes": 1024, "tensorSha256": "8fb278e603920c74ed391efb69569542ea579806fc2c73b65e240f83bf1f1130", "rangeHttpStatus": 206, "contentRange": "bytes 628608-629631/5363339264", "valueCount": 256, "min": 3.68896484375, "max": 4.0338897705078125, "spread": 0.3449249267578125, "mean": 3.9844262897968292, "standardDeviation": 0.04810790123536426, "fp32Distinct": 235, "bf16Distinct": 16, "bf16Min": 3.6875, "bf16Max": 4.03125, "meanAbsoluteQuantizationError": 0.0059039779007434845, "maxAbsoluteQuantizationError": 0.01534891128540039, "zeroLogitTop8SetDifference": 8 }, { "layer": 31, "tensor": "model.layers.31.mlp.gate.e_score_correction_bias", "shard": "model-00044-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105896, "headerSha256": "c6872cd77e3ff3149a95126508afd56625a3135fb78eb70c9252f13efbe03ca5", "tensorByteRange": [ 422320, 423343 ], "tensorBytes": 1024, "tensorSha256": "9c6484e08e10f663b3ab1912b27f379c25c002f1557fbc06541689111daef9fe", "rangeHttpStatus": 206, "contentRange": "bytes 422320-423343/5366353776", "valueCount": 256, "min": 3.9608144760131836, "max": 4.341724395751953, "spread": 0.38090991973876953, "mean": 4.292792621999979, "standardDeviation": 0.05136393300190469, "fp32Distinct": 240, "bf16Distinct": 10, "bf16Min": 3.953125, "bf16Max": 4.34375, "meanAbsoluteQuantizationError": 0.008535653352737427, "maxAbsoluteQuantizationError": 0.015506744384765625, "zeroLogitTop8SetDifference": 7 }, { "layer": 32, "tensor": "model.layers.32.mlp.gate.e_score_correction_bias", "shard": "model-00046-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106040, "headerSha256": "641e5f7ba936134688b2afe7774cd6e4f011125c9915ae1c59f82bf8b8883a14", "tensorByteRange": [ 213568, 214591 ], "tensorBytes": 1024, "tensorSha256": "1097f6fbf81795cbc416940189174069c7a38d7fff47bd6213fd38d1514a178b", "rangeHttpStatus": 206, "contentRange": "bytes 213568-214591/5366353920", "valueCount": 256, "min": 6.015722274780273, "max": 6.408608436584473, "spread": 0.3928861618041992, "mean": 6.356187757104635, "standardDeviation": 0.05868773635246458, "fp32Distinct": 235, "bf16Distinct": 11, "bf16Min": 6.03125, "bf16Max": 6.40625, "meanAbsoluteQuantizationError": 0.007736179977655411, "maxAbsoluteQuantizationError": 0.015616416931152344, "zeroLogitTop8SetDifference": 6 }, { "layer": 33, "tensor": "model.layers.33.mlp.gate.e_score_correction_bias", "shard": "model-00047-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105616, "headerSha256": "3ef0ec6671ab43278b2ed7e629f2ae08e42e79b060d0d423894b41bd1bc79521", "tensorByteRange": [ 1312920, 1313943 ], "tensorBytes": 1024, "tensorSha256": "9eed8b613aade062ef7ce223060c882a9762f9e26472f37f3858a8e4599bbaf6", "rangeHttpStatus": 206, "contentRange": "bytes 1312920-1313943/5366353496", "valueCount": 256, "min": 4.7284345626831055, "max": 5.103320121765137, "spread": 0.37488555908203125, "mean": 5.041470346972346, "standardDeviation": 0.055001823364684066, "fp32Distinct": 240, "bf16Distinct": 10, "bf16Min": 4.71875, "bf16Max": 5.09375, "meanAbsoluteQuantizationError": 0.008116750046610832, "maxAbsoluteQuantizationError": 0.015615463256835938, "zeroLogitTop8SetDifference": 8 }, { "layer": 34, "tensor": "model.layers.34.mlp.gate.e_score_correction_bias", "shard": "model-00049-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106200, "headerSha256": "5b58e26a404127b9852e5ade264dfcb8b010664d4f8c0314ad94a7ec230ad0e3", "tensorByteRange": [ 1104608, 1105631 ], "tensorBytes": 1024, "tensorSha256": "d16804b6b5c8e4693a556633d4a9d659f3e642dc5718adaae2802a204e33aca8", "rangeHttpStatus": 206, "contentRange": "bytes 1104608-1105631/5363339104", "valueCount": 256, "min": 5.12522029876709, "max": 5.543081283569336, "spread": 0.4178609848022461, "mean": 5.477707354351878, "standardDeviation": 0.0642430213089266, "fp32Distinct": 238, "bf16Distinct": 10, "bf16Min": 5.125, "bf16Max": 5.53125, "meanAbsoluteQuantizationError": 0.007306938990950584, "maxAbsoluteQuantizationError": 0.015582084655761719, "zeroLogitTop8SetDifference": 8 }, { "layer": 35, "tensor": "model.layers.35.mlp.gate.e_score_correction_bias", "shard": "model-00051-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105632, "headerSha256": "9cac2749078e434766979e5526c1bb9a15c2b8e48ba2055a5a9813864f1cc99f", "tensorByteRange": [ 898216, 899239 ], "tensorBytes": 1024, "tensorSha256": "f252465e5f4ff036da3a43c14e04289a8a837f7c7caa63b9bf61a07bf0422f6b", "rangeHttpStatus": 206, "contentRange": "bytes 898216-899239/5366353512", "valueCount": 256, "min": 5.480034828186035, "max": 5.912888526916504, "spread": 0.43285369873046875, "mean": 5.84933459572494, "standardDeviation": 0.06719110068360684, "fp32Distinct": 237, "bf16Distinct": 12, "bf16Min": 5.46875, "bf16Max": 5.90625, "meanAbsoluteQuantizationError": 0.007363399490714073, "maxAbsoluteQuantizationError": 0.01556396484375, "zeroLogitTop8SetDifference": 7 }, { "layer": 36, "tensor": "model.layers.36.mlp.gate.e_score_correction_bias", "shard": "model-00053-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105728, "headerSha256": "02f4f42ac18e565c1c8d9f4ef506fc4b026e4e95d2ce6e4a3cb40b0b5e3df2b8", "tensorByteRange": [ 689416, 690439 ], "tensorBytes": 1024, "tensorSha256": "8d0b9596512efc8c6e600d3ea21ab0f67c8d9a510132b6d63d8ce24afaa40439", "rangeHttpStatus": 206, "contentRange": "bytes 689416-690439/5366353608", "valueCount": 256, "min": 4.31663703918457, "max": 4.776505470275879, "spread": 0.4598684310913086, "mean": 4.7179800141602755, "standardDeviation": 0.06694534647451324, "fp32Distinct": 233, "bf16Distinct": 13, "bf16Min": 4.3125, "bf16Max": 4.78125, "meanAbsoluteQuantizationError": 0.008786549791693687, "maxAbsoluteQuantizationError": 0.015520095825195312, "zeroLogitTop8SetDifference": 8 }, { "layer": 37, "tensor": "model.layers.37.mlp.gate.e_score_correction_bias", "shard": "model-00055-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105864, "headerSha256": "9d5cecac6007446cf9947ffe5d2ad21a784f3f4beebe7b14c3fb12b96b3d6147", "tensorByteRange": [ 480656, 481679 ], "tensorBytes": 1024, "tensorSha256": "2992d0b8ff22654f381fa8ebdf6413ac97062a0ce63f49f3060e8bef175539d6", "rangeHttpStatus": 206, "contentRange": "bytes 480656-481679/5366353744", "valueCount": 256, "min": 5.257145881652832, "max": 5.608041763305664, "spread": 0.35089588165283203, "mean": 5.548734096810222, "standardDeviation": 0.06774260311923073, "fp32Distinct": 238, "bf16Distinct": 12, "bf16Min": 5.25, "bf16Max": 5.59375, "meanAbsoluteQuantizationError": 0.007226267829537392, "maxAbsoluteQuantizationError": 0.015562057495117188, "zeroLogitTop8SetDifference": 6 }, { "layer": 38, "tensor": "model.layers.38.mlp.gate.e_score_correction_bias", "shard": "model-00057-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106592, "headerSha256": "cefedaf9781aec46c99976be93c58166a2b4e3cba115390a0f454a7f73625e33", "tensorByteRange": [ 272488, 273511 ], "tensorBytes": 1024, "tensorSha256": "33dfe33620069c360f9a3ce3b0ad8aa252016769b6051ef09d0e15036d39b7ff", "rangeHttpStatus": 206, "contentRange": "bytes 272488-273511/5363339496", "valueCount": 256, "min": 4.410601615905762, "max": 4.906434059143066, "spread": 0.4958324432373047, "mean": 4.843597872182727, "standardDeviation": 0.07845141867278854, "fp32Distinct": 237, "bf16Distinct": 13, "bf16Min": 4.40625, "bf16Max": 4.90625, "meanAbsoluteQuantizationError": 0.007996032014489174, "maxAbsoluteQuantizationError": 0.015552520751953125, "zeroLogitTop8SetDifference": 7 }, { "layer": 39, "tensor": "model.layers.39.mlp.gate.e_score_correction_bias", "shard": "model-00058-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105808, "headerSha256": "c12f89e3a1b6ef302be9ccae279d4e8608c8a0136304837897c939c7be29a37f", "tensorByteRange": [ 1374552, 1375575 ], "tensorBytes": 1024, "tensorSha256": "2592c31b6cc98503ba30794ec6c13f35ea2cdee78d9331f0a5ed7b5a1e899884", "rangeHttpStatus": 206, "contentRange": "bytes 1374552-1375575/5357950744", "valueCount": 256, "min": 5.983743667602539, "max": 6.378626823425293, "spread": 0.3948831558227539, "mean": 6.31624841876328, "standardDeviation": 0.07350414249414026, "fp32Distinct": 230, "bf16Distinct": 14, "bf16Min": 5.96875, "bf16Max": 6.375, "meanAbsoluteQuantizationError": 0.007859738543629646, "maxAbsoluteQuantizationError": 0.015405654907226562, "zeroLogitTop8SetDifference": 7 }, { "layer": 40, "tensor": "model.layers.40.mlp.gate.e_score_correction_bias", "shard": "model-00062-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105624, "headerSha256": "223b2acdc251053c81433b4a478b8987db2e7ce01f37c64232f03d71a54b56ab", "tensorByteRange": [ 959648, 960671 ], "tensorBytes": 1024, "tensorSha256": "64c6d43b2fc3ee6616bf1174384054574c44db835a4ffb4abf5c400212f0e700", "rangeHttpStatus": 206, "contentRange": "bytes 959648-960671/5366353504", "valueCount": 256, "min": 7.851768493652344, "max": 8.29575252532959, "spread": 0.4439840316772461, "mean": 8.236589845269918, "standardDeviation": 0.07068335867638376, "fp32Distinct": 225, "bf16Distinct": 8, "bf16Min": 7.84375, "bf16Max": 8.3125, "meanAbsoluteQuantizationError": 0.01911274716258049, "maxAbsoluteQuantizationError": 0.03125, "zeroLogitTop8SetDifference": 7 }, { "layer": 41, "tensor": "model.layers.41.mlp.gate.e_score_correction_bias", "shard": "model-00064-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105696, "headerSha256": "cfa773ee02dffea70b560b73f88f83047b6f080b65baa842a996abac0d69d224", "tensorByteRange": [ 750824, 751847 ], "tensorBytes": 1024, "tensorSha256": "b168f0d57e2bb7dd59ac3d9577bbd9bc834509f62f9b30ef06a03105755830c7", "rangeHttpStatus": 206, "contentRange": "bytes 750824-751847/5366353576", "valueCount": 256, "min": 6.2715959548950195, "max": 6.698460578918457, "spread": 0.4268646240234375, "mean": 6.642226027324796, "standardDeviation": 0.07548607927431586, "fp32Distinct": 225, "bf16Distinct": 13, "bf16Min": 6.28125, "bf16Max": 6.6875, "meanAbsoluteQuantizationError": 0.006688373163342476, "maxAbsoluteQuantizationError": 0.015504837036132812, "zeroLogitTop8SetDifference": 8 }, { "layer": 42, "tensor": "model.layers.42.mlp.gate.e_score_correction_bias", "shard": "model-00066-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106416, "headerSha256": "35ae2ede04db240c4ab0ae56c7e42d45a96d47cb46a1eb0ca8f0cdd5c9b75b22", "tensorByteRange": [ 542648, 543671 ], "tensorBytes": 1024, "tensorSha256": "b40efd174dd4756d63ab462ab49daf65b3ab4ce994b13b5a6622f1b825caf5b9", "rangeHttpStatus": 206, "contentRange": "bytes 542648-543671/5363339320", "valueCount": 256, "min": 9.498008728027344, "max": 9.937845230102539, "spread": 0.4398365020751953, "mean": 9.88309109956026, "standardDeviation": 0.07195092136381114, "fp32Distinct": 225, "bf16Distinct": 8, "bf16Min": 9.5, "bf16Max": 9.9375, "meanAbsoluteQuantizationError": 0.015273191034793854, "maxAbsoluteQuantizationError": 0.030942916870117188, "zeroLogitTop8SetDifference": 8 }, { "layer": 43, "tensor": "model.layers.43.mlp.gate.e_score_correction_bias", "shard": "model-00068-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105968, "headerSha256": "9848648a4673c8533951df63c5a6d9e8eea307dc1e075e4b90da36bddf1157d4", "tensorByteRange": [ 336376, 337399 ], "tensorBytes": 1024, "tensorSha256": "9965776c660b7de79ee21674bc7aa750f7f6d31327b495e25efb8cc76ec3b491", "rangeHttpStatus": 206, "contentRange": "bytes 336376-337399/5366353848", "valueCount": 256, "min": 5.090263366699219, "max": 5.471126556396484, "spread": 0.3808631896972656, "mean": 5.418202552944422, "standardDeviation": 0.06425906877692787, "fp32Distinct": 227, "bf16Distinct": 11, "bf16Min": 5.09375, "bf16Max": 5.46875, "meanAbsoluteQuantizationError": 0.008914917707443237, "maxAbsoluteQuantizationError": 0.01561737060546875, "zeroLogitTop8SetDifference": 7 }, { "layer": 44, "tensor": "model.layers.44.mlp.gate.e_score_correction_bias", "shard": "model-00070-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106112, "headerSha256": "6bfb373a3e6ce708a59e9f1daf31dfda05629b7bbf4e6a20db4702629fd49fef", "tensorByteRange": [ 127624, 128647 ], "tensorBytes": 1024, "tensorSha256": "433ca74caf9e783d76cb5c7c381f940a6370294eef35089e7825b3b3267bd323", "rangeHttpStatus": 206, "contentRange": "bytes 127624-128647/5366353992", "valueCount": 256, "min": 5.750890731811523, "max": 6.147760391235352, "spread": 0.3968696594238281, "mean": 6.093448961153626, "standardDeviation": 0.06703125061528707, "fp32Distinct": 220, "bf16Distinct": 13, "bf16Min": 5.75, "bf16Max": 6.15625, "meanAbsoluteQuantizationError": 0.008207255974411964, "maxAbsoluteQuantizationError": 0.015547752380371094, "zeroLogitTop8SetDifference": 5 }, { "layer": 45, "tensor": "model.layers.45.mlp.gate.e_score_correction_bias", "shard": "model-00071-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105608, "headerSha256": "b58e189499f71cdd06e79d17f09fa2ad61b2f4a5669c61885d2f8f680110283a", "tensorByteRange": [ 1226896, 1227919 ], "tensorBytes": 1024, "tensorSha256": "0606b768b7dc73fc6a1510070be1c4a06e025d5d2dde11427e79b7b24090cb10", "rangeHttpStatus": 206, "contentRange": "bytes 1226896-1227919/5366353488", "valueCount": 256, "min": 6.820390224456787, "max": 7.338102340698242, "spread": 0.5177121162414551, "mean": 7.284339310601354, "standardDeviation": 0.07150897153036215, "fp32Distinct": 217, "bf16Distinct": 14, "bf16Min": 6.8125, "bf16Max": 7.34375, "meanAbsoluteQuantizationError": 0.00851668231189251, "maxAbsoluteQuantizationError": 0.015608787536621094, "zeroLogitTop8SetDifference": 6 }, { "layer": 46, "tensor": "model.layers.46.mlp.gate.e_score_correction_bias", "shard": "model-00073-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106208, "headerSha256": "5ac88ebbebf161bf9135f47fe6cbaaf4ceca88f92b46a64a2d102782528856d3", "tensorByteRange": [ 1018600, 1019623 ], "tensorBytes": 1024, "tensorSha256": "ef5a1d7c86cb647a9a7e3af5aa7ec052e7060648c8e73d1b3dd442da50e2bf75", "rangeHttpStatus": 206, "contentRange": "bytes 1018600-1019623/5363339112", "valueCount": 256, "min": 8.590546607971191, "max": 9.217239379882812, "spread": 0.6266927719116211, "mean": 9.169565096497536, "standardDeviation": 0.07166289587911094, "fp32Distinct": 212, "bf16Distinct": 7, "bf16Min": 8.5625, "bf16Max": 9.1875, "meanAbsoluteQuantizationError": 0.017439275979995728, "maxAbsoluteQuantizationError": 0.03112030029296875, "zeroLogitTop8SetDifference": 8 }, { "layer": 47, "tensor": "model.layers.47.mlp.gate.e_score_correction_bias", "shard": "model-00075-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105648, "headerSha256": "aa503b115df2977010d076147577bb251e224c3ae392b998528b5337d3062442", "tensorByteRange": [ 812216, 813239 ], "tensorBytes": 1024, "tensorSha256": "aa720ba3742e1ed27b47f188015c80f07cd89e5c8c13493102b00a32620399e1", "rangeHttpStatus": 206, "contentRange": "bytes 812216-813239/5366353528", "valueCount": 256, "min": 6.0587921142578125, "max": 6.645472526550293, "spread": 0.5866804122924805, "mean": 6.5932945497334, "standardDeviation": 0.07143207951061967, "fp32Distinct": 220, "bf16Distinct": 15, "bf16Min": 6.0625, "bf16Max": 6.65625, "meanAbsoluteQuantizationError": 0.007830895483493805, "maxAbsoluteQuantizationError": 0.015517234802246094, "zeroLogitTop8SetDifference": 7 }, { "layer": 48, "tensor": "model.layers.48.mlp.gate.e_score_correction_bias", "shard": "model-00077-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105792, "headerSha256": "08b43435aa593b4c69c0e30e48785611635d3def5371efe8ba4d110b664f3c8a", "tensorByteRange": [ 603464, 604487 ], "tensorBytes": 1024, "tensorSha256": "e07a3ba7b83144d43cb9fe668798cee42ffb82d79978a7b718b8c0a95f2bb9e3", "rangeHttpStatus": 206, "contentRange": "bytes 603464-604487/5366353672", "valueCount": 256, "min": 9.0234375, "max": 9.463132858276367, "spread": 0.4396953582763672, "mean": 9.419414561241865, "standardDeviation": 0.06655449656213692, "fp32Distinct": 214, "bf16Distinct": 7, "bf16Min": 9, "bf16Max": 9.4375, "meanAbsoluteQuantizationError": 0.01336711272597313, "maxAbsoluteQuantizationError": 0.031131744384765625, "zeroLogitTop8SetDifference": 7 }, { "layer": 49, "tensor": "model.layers.49.mlp.gate.e_score_correction_bias", "shard": "model-00079-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105288, "headerSha256": "7b09be4b555734a0178da13b4de63f9832eb3d2d7210cacfd2858ce52bd21ff9", "tensorByteRange": [ 394064, 395087 ], "tensorBytes": 1024, "tensorSha256": "5fce54a03ef464dcbbe9c9aab63242c6338ca4f91870cee0d3d6e0377864f799", "rangeHttpStatus": 206, "contentRange": "bytes 394064-395087/5366353168", "valueCount": 256, "min": 8.221697807312012, "max": 8.53561019897461, "spread": 0.31391239166259766, "mean": 8.495769917964935, "standardDeviation": 0.055094847637572525, "fp32Distinct": 212, "bf16Distinct": 6, "bf16Min": 8.25, "bf16Max": 8.5625, "meanAbsoluteQuantizationError": 0.017932847142219543, "maxAbsoluteQuantizationError": 0.030889511108398438, "zeroLogitTop8SetDifference": 5 }, { "layer": 50, "tensor": "model.layers.50.mlp.gate.e_score_correction_bias", "shard": "model-00082-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106200, "headerSha256": "24cc512972ccce30ec7e31987593a23e8052754f9915ab9857aa8ada7b11dccf", "tensorByteRange": [ 1285856, 1286879 ], "tensorBytes": 1024, "tensorSha256": "df6f38dd02d86cc49e2b59d3852ca16e7c017e9f32fbf6964bd00a5fcc9a3958", "rangeHttpStatus": 206, "contentRange": "bytes 1285856-1286879/5363339104", "valueCount": 256, "min": 9.047330856323242, "max": 9.421131134033203, "spread": 0.37380027770996094, "mean": 9.38666221871972, "standardDeviation": 0.045545782083579664, "fp32Distinct": 210, "bf16Distinct": 6, "bf16Min": 9.0625, "bf16Max": 9.4375, "meanAbsoluteQuantizationError": 0.021220479160547256, "maxAbsoluteQuantizationError": 0.031152725219726562, "zeroLogitTop8SetDifference": 8 }, { "layer": 51, "tensor": "model.layers.51.mlp.gate.e_score_correction_bias", "shard": "model-00084-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105616, "headerSha256": "c5812ad0a01405671a3f974b1c22b3a132dad50158d28b2f14e4444ad7987ef7", "tensorByteRange": [ 1079448, 1080471 ], "tensorBytes": 1024, "tensorSha256": "f25f7bcc0a15231c44a2898b1e3baec0664c4ef05403723b9999a6259279c008", "rangeHttpStatus": 206, "contentRange": "bytes 1079448-1080471/5366353496", "valueCount": 256, "min": 9.574926376342773, "max": 9.994817733764648, "spread": 0.419891357421875, "mean": 9.964124862104654, "standardDeviation": 0.050414055997246875, "fp32Distinct": 204, "bf16Distinct": 6, "bf16Min": 9.5625, "bf16Max": 10, "meanAbsoluteQuantizationError": 0.016226213425397873, "maxAbsoluteQuantizationError": 0.031147003173828125, "zeroLogitTop8SetDifference": 8 }, { "layer": 52, "tensor": "model.layers.52.mlp.gate.e_score_correction_bias", "shard": "model-00086-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105632, "headerSha256": "0bb78438cf49714bffbb9e233bfae8e61b50682e620cf8039299d62f9ccd4024", "tensorByteRange": [ 870568, 871591 ], "tensorBytes": 1024, "tensorSha256": "0898bdd123b624a7037743ab5eff41126578ad610e3a07c6f3ef0539648b7956", "rangeHttpStatus": 206, "contentRange": "bytes 870568-871591/5366353512", "valueCount": 256, "min": 8.832311630249023, "max": 9.062313079833984, "spread": 0.23000144958496094, "mean": 9.034025553613901, "standardDeviation": 0.03464267465855478, "fp32Distinct": 204, "bf16Distinct": 5, "bf16Min": 8.8125, "bf16Max": 9.0625, "meanAbsoluteQuantizationError": 0.013770584017038345, "maxAbsoluteQuantizationError": 0.03036975860595703, "zeroLogitTop8SetDifference": 8 }, { "layer": 53, "tensor": "model.layers.53.mlp.gate.e_score_correction_bias", "shard": "model-00088-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105752, "headerSha256": "c99eab2a5186acb315022a8b476275c556fac6d836cbce2f9aa180324ebf1136", "tensorByteRange": [ 661792, 662815 ], "tensorBytes": 1024, "tensorSha256": "6a0de02f81a3ca36f4c8b8fa87742daff81ad79ffca280ff23b4669c28dbd082", "rangeHttpStatus": 206, "contentRange": "bytes 661792-662815/5366353632", "valueCount": 256, "min": 8.723367691040039, "max": 9.062324523925781, "spread": 0.3389568328857422, "mean": 9.03482074663043, "standardDeviation": 0.042241537179600455, "fp32Distinct": 189, "bf16Distinct": 6, "bf16Min": 8.75, "bf16Max": 9.0625, "meanAbsoluteQuantizationError": 0.013552282005548477, "maxAbsoluteQuantizationError": 0.03118133544921875, "zeroLogitTop8SetDifference": 8 }, { "layer": 54, "tensor": "model.layers.54.mlp.gate.e_score_correction_bias", "shard": "model-00090-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106472, "headerSha256": "ba089e67d9f1dff0fc6071e33f452c1f262b809472454db677c257061ed2332a", "tensorByteRange": [ 453616, 454639 ], "tensorBytes": 1024, "tensorSha256": "9415c45381b70df5013eb01577b935065fc7129747523ff6a5de7f46d834ecb7", "rangeHttpStatus": 206, "contentRange": "bytes 453616-454639/5363339376", "valueCount": 256, "min": 8.786428451538086, "max": 9.429126739501953, "spread": 0.6426982879638672, "mean": 9.398599572479725, "standardDeviation": 0.06101513716608349, "fp32Distinct": 185, "bf16Distinct": 8, "bf16Min": 8.8125, "bf16Max": 9.4375, "meanAbsoluteQuantizationError": 0.018075622618198395, "maxAbsoluteQuantizationError": 0.03115081787109375, "zeroLogitTop8SetDifference": 8 }, { "layer": 55, "tensor": "model.layers.55.mlp.gate.e_score_correction_bias", "shard": "model-00092-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106024, "headerSha256": "5c75fa757fb050c4fad97924d3eed3c83ffeb25f1f539c1059616b717185e77b", "tensorByteRange": [ 247344, 248367 ], "tensorBytes": 1024, "tensorSha256": "e4a9c6b50225d64ef482480b3755ad4aa94a4832f3a2aa03e963c0a7588a8789", "rangeHttpStatus": 206, "contentRange": "bytes 247344-248367/5366353904", "valueCount": 256, "min": 7.994711875915527, "max": 8.286730766296387, "spread": 0.2920188903808594, "mean": 8.257194951176643, "standardDeviation": 0.03774304809284942, "fp32Distinct": 183, "bf16Distinct": 6, "bf16Min": 8, "bf16Max": 8.3125, "meanAbsoluteQuantizationError": 0.020398393273353577, "maxAbsoluteQuantizationError": 0.031218528747558594, "zeroLogitTop8SetDifference": 3 }, { "layer": 56, "tensor": "model.layers.56.mlp.gate.e_score_correction_bias", "shard": "model-00093-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105608, "headerSha256": "4bef5fd9a08ee21882ecb76cfce6bcf134f115af03a75824e831f177353c037c", "tensorByteRange": [ 1346704, 1347727 ], "tensorBytes": 1024, "tensorSha256": "8ab5a27d2eea11f76a6c776a05ac7b39bae621ab912e2b3ce8f3475cee049e35", "rangeHttpStatus": 206, "contentRange": "bytes 1346704-1347727/5366353488", "valueCount": 256, "min": 9.888904571533203, "max": 10.189737319946289, "spread": 0.30083274841308594, "mean": 10.164271153509617, "standardDeviation": 0.04002725662172432, "fp32Distinct": 176, "bf16Distinct": 5, "bf16Min": 9.875, "bf16Max": 10.1875, "meanAbsoluteQuantizationError": 0.01088552176952362, "maxAbsoluteQuantizationError": 0.031232833862304688, "zeroLogitTop8SetDifference": 7 }, { "layer": 57, "tensor": "model.layers.57.mlp.gate.e_score_correction_bias", "shard": "model-00095-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105608, "headerSha256": "b6b52e9ed3dd2b87b4acada6318504b84287accafc6bdf82426645d275e426c1", "tensorByteRange": [ 1137808, 1138831 ], "tensorBytes": 1024, "tensorSha256": "9f08571053bc9a527bc4682d0a9502394336093d672b530170dc8f7d50248e2f", "rangeHttpStatus": 206, "contentRange": "bytes 1137808-1138831/5366353488", "valueCount": 256, "min": 11.02508544921875, "max": 11.292142868041992, "spread": 0.2670574188232422, "mean": 11.266072932630777, "standardDeviation": 0.0352182599236895, "fp32Distinct": 179, "bf16Distinct": 6, "bf16Min": 11, "bf16Max": 11.3125, "meanAbsoluteQuantizationError": 0.02154945209622383, "maxAbsoluteQuantizationError": 0.03115081787109375, "zeroLogitTop8SetDifference": 8 }, { "layer": 58, "tensor": "model.layers.58.mlp.gate.e_score_correction_bias", "shard": "model-00097-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106208, "headerSha256": "c1232b2a67fe0dfca39f6f0c931a4e1a3c9e16956d3d8ff0cf6f35062bc7d5ae", "tensorByteRange": [ 929512, 930535 ], "tensorBytes": 1024, "tensorSha256": "0dcd01df0871b7ea812917e8371021e231b8ee66646d4d623209b5257fb4055d", "rangeHttpStatus": 206, "contentRange": "bytes 929512-930535/5363339112", "valueCount": 256, "min": 7.459988594055176, "max": 7.76688289642334, "spread": 0.30689430236816406, "mean": 7.74039557017386, "standardDeviation": 0.0373759122653254, "fp32Distinct": 191, "bf16Distinct": 9, "bf16Min": 7.46875, "bf16Max": 7.78125, "meanAbsoluteQuantizationError": 0.008197678253054619, "maxAbsoluteQuantizationError": 0.015570640563964844, "zeroLogitTop8SetDifference": 2 }, { "layer": 59, "tensor": "model.layers.59.mlp.gate.e_score_correction_bias", "shard": "model-00099-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105288, "headerSha256": "5e3e3ef6b7dc5ff78f3c766982e205578183750bdd4d3b74c9361ca4226fde12", "tensorByteRange": [ 722768, 723791 ], "tensorBytes": 1024, "tensorSha256": "4079aa06fab5e721e751ca5bf446c43f784eb9c4904391cd23ce1fbd5e07c79e", "rangeHttpStatus": 206, "contentRange": "bytes 722768-723791/5366353168", "valueCount": 256, "min": 9.47209358215332, "max": 9.806940078735352, "spread": 0.33484649658203125, "mean": 9.781522776931524, "standardDeviation": 0.03914750421483456, "fp32Distinct": 183, "bf16Distinct": 5, "bf16Min": 9.5, "bf16Max": 9.8125, "meanAbsoluteQuantizationError": 0.015133555978536606, "maxAbsoluteQuantizationError": 0.03106689453125, "zeroLogitTop8SetDifference": 8 }, { "layer": 60, "tensor": "model.layers.60.mlp.gate.e_score_correction_bias", "shard": "model-00103-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105976, "headerSha256": "3f7a3668fb46cff2a875aa155e731d9f8c292f1b2b16c17650ccb35f44592c3a", "tensorByteRange": [ 308736, 309759 ], "tensorBytes": 1024, "tensorSha256": "be29f2e518f92f0bc88cd5e37cad691703fe0f47266698aa2c0939b78f33b0e4", "rangeHttpStatus": 206, "contentRange": "bytes 308736-309759/5366353856", "valueCount": 256, "min": 10.285585403442383, "max": 10.471578598022461, "spread": 0.18599319458007812, "mean": 10.44658637791872, "standardDeviation": 0.031100147124989035, "fp32Distinct": 186, "bf16Distinct": 4, "bf16Min": 10.3125, "bf16Max": 10.5, "meanAbsoluteQuantizationError": 0.01999013125896454, "maxAbsoluteQuantizationError": 0.0310821533203125, "zeroLogitTop8SetDifference": 1 }, { "layer": 61, "tensor": "model.layers.61.mlp.gate.e_score_correction_bias", "shard": "model-00104-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 107264, "headerSha256": "3f6a55faf3c86a850bd261f17da50f17ceedb96d97c065f0e3615c94fd6e7017", "tensorByteRange": [ 1409800, 1410823 ], "tensorBytes": 1024, "tensorSha256": "f43b95c79a753aa78bc479c514411da456c6aa81028caa21877109cb139e5e10", "rangeHttpStatus": 206, "contentRange": "bytes 1409800-1410823/5364883208", "valueCount": 256, "min": 11.250185012817383, "max": 11.597997665405273, "spread": 0.3478126525878906, "mean": 11.572012569755316, "standardDeviation": 0.041116391410031906, "fp32Distinct": 190, "bf16Distinct": 7, "bf16Min": 11.25, "bf16Max": 11.625, "meanAbsoluteQuantizationError": 0.020898904651403427, "maxAbsoluteQuantizationError": 0.031007766723632812, "zeroLogitTop8SetDifference": 6 }, { "layer": 62, "tensor": "model.layers.62.mlp.gate.e_score_correction_bias", "shard": "model-00106-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106200, "headerSha256": "66e96f56fde5c0980cf89e56fb1b0514410dd23507d207001d714a53aa12e167", "tensorByteRange": [ 1199840, 1200863 ], "tensorBytes": 1024, "tensorSha256": "2c94000bb3dc2492f32d10287635738421866117146a0ee370983b8cb0b4276b", "rangeHttpStatus": 206, "contentRange": "bytes 1199840-1200863/5363339104", "valueCount": 256, "min": 9.532155990600586, "max": 9.697067260742188, "spread": 0.16491127014160156, "mean": 9.674441948533058, "standardDeviation": 0.02878025088624029, "fp32Distinct": 172, "bf16Distinct": 3, "bf16Min": 9.5625, "bf16Max": 9.6875, "meanAbsoluteQuantizationError": 0.009173326194286346, "maxAbsoluteQuantizationError": 0.031072616577148438, "zeroLogitTop8SetDifference": 8 }, { "layer": 63, "tensor": "model.layers.63.mlp.gate.e_score_correction_bias", "shard": "model-00108-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105624, "headerSha256": "693d32ea11a3f9a199fe77dace6e5482430da7554c84e9648d507e025720c342", "tensorByteRange": [ 993440, 994463 ], "tensorBytes": 1024, "tensorSha256": "35ea76545f95785a12acbc8291549699d1f0ab5352eaac4ef4f5052b350ddfb8", "rangeHttpStatus": 206, "contentRange": "bytes 993440-994463/5366353504", "valueCount": 256, "min": 12.973209381103516, "max": 13.287189483642578, "spread": 0.3139801025390625, "mean": 13.26252031326294, "standardDeviation": 0.03785927920038606, "fp32Distinct": 186, "bf16Distinct": 6, "bf16Min": 13, "bf16Max": 13.3125, "meanAbsoluteQuantizationError": 0.021850720047950745, "maxAbsoluteQuantizationError": 0.031198501586914062, "zeroLogitTop8SetDifference": 6 }, { "layer": 64, "tensor": "model.layers.64.mlp.gate.e_score_correction_bias", "shard": "model-00110-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105672, "headerSha256": "f7b4cfa6a244b5a9f6b2d42845da02d0fb22d64097e4162d0458ef57acafe695", "tensorByteRange": [ 784592, 785615 ], "tensorBytes": 1024, "tensorSha256": "0f5ae4abb0d8300de214aec29ceef2a2eb1608a4a2de6021179ca07668b340a1", "rangeHttpStatus": 206, "contentRange": "bytes 784592-785615/5366353552", "valueCount": 256, "min": 11.12826156616211, "max": 11.31718635559082, "spread": 0.18892478942871094, "mean": 11.292042881250381, "standardDeviation": 0.028851026041330493, "fp32Distinct": 196, "bf16Distinct": 4, "bf16Min": 11.125, "bf16Max": 11.3125, "meanAbsoluteQuantizationError": 0.010759249329566956, "maxAbsoluteQuantizationError": 0.031217575073242188, "zeroLogitTop8SetDifference": 8 }, { "layer": 65, "tensor": "model.layers.65.mlp.gate.e_score_correction_bias", "shard": "model-00112-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105808, "headerSha256": "23d05dbe856cec988e52a1d06a3fa0ac3c0a4706d624d026d2bfeba564b11999", "tensorByteRange": [ 575832, 576855 ], "tensorBytes": 1024, "tensorSha256": "9eb32689d59459a82b4b7759f574a5f6bcc9bac1a44884d359a83250587f644e", "rangeHttpStatus": 206, "contentRange": "bytes 575832-576855/5366353688", "valueCount": 256, "min": 13.029184341430664, "max": 13.306112289428711, "spread": 0.2769279479980469, "mean": 13.279166478663683, "standardDeviation": 0.03347844359040279, "fp32Distinct": 184, "bf16Distinct": 5, "bf16Min": 13, "bf16Max": 13.3125, "meanAbsoluteQuantizationError": 0.018129367381334305, "maxAbsoluteQuantizationError": 0.03113555908203125, "zeroLogitTop8SetDifference": 8 }, { "layer": 66, "tensor": "model.layers.66.mlp.gate.e_score_correction_bias", "shard": "model-00114-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106528, "headerSha256": "91ad8776e44c0e4bcd963a7ad0ec4fa3974b9e9bbb93f76748e868102742a377", "tensorByteRange": [ 367656, 368679 ], "tensorBytes": 1024, "tensorSha256": "c8d54d8a84a33283d142ddcbf98d3fc71c97d96ec158678402258e1a20604a66", "rangeHttpStatus": 206, "contentRange": "bytes 367656-368679/5363339432", "valueCount": 256, "min": 11.995874404907227, "max": 12.430606842041016, "spread": 0.43473243713378906, "mean": 12.40172153711319, "standardDeviation": 0.046383939325787646, "fp32Distinct": 178, "bf16Distinct": 7, "bf16Min": 12, "bf16Max": 12.4375, "meanAbsoluteQuantizationError": 0.017378106713294983, "maxAbsoluteQuantizationError": 0.030895233154296875, "zeroLogitTop8SetDifference": 7 }, { "layer": 67, "tensor": "model.layers.67.mlp.gate.e_score_correction_bias", "shard": "model-00116-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106096, "headerSha256": "4ccc898397e5120ecd0d01b43a0b8c43b2e7ee34cd9ea74f679a37051a2ea87f", "tensorByteRange": [ 161400, 162423 ], "tensorBytes": 1024, "tensorSha256": "869731624b19ceab0cdc1f36aef85fb28dc24b61f0bd6f535ce644df264dea4c", "rangeHttpStatus": 206, "contentRange": "bytes 161400-162423/5366353976", "valueCount": 256, "min": 12.709373474121094, "max": 13.08515739440918, "spread": 0.37578392028808594, "mean": 13.056553814560175, "standardDeviation": 0.03856610312209759, "fp32Distinct": 200, "bf16Distinct": 6, "bf16Min": 12.6875, "bf16Max": 13.0625, "meanAbsoluteQuantizationError": 0.012780312448740005, "maxAbsoluteQuantizationError": 0.031169891357421875, "zeroLogitTop8SetDifference": 8 }, { "layer": 68, "tensor": "model.layers.68.mlp.gate.e_score_correction_bias", "shard": "model-00117-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105608, "headerSha256": "d3c865fee37f7287ca8c8e1170ddea6af247e85a50aed91902013712cb323001", "tensorByteRange": [ 1260688, 1261711 ], "tensorBytes": 1024, "tensorSha256": "cc70198f190b2c665f1d162eab0328e974ac9538ac0538cfcde50b554853bdbd", "rangeHttpStatus": 206, "contentRange": "bytes 1260688-1261711/5366353488", "valueCount": 256, "min": 12.7413330078125, "max": 12.984230041503906, "spread": 0.24289703369140625, "mean": 12.953461766242981, "standardDeviation": 0.03446610197623668, "fp32Distinct": 206, "bf16Distinct": 5, "bf16Min": 12.75, "bf16Max": 13, "meanAbsoluteQuantizationError": 0.02096664160490036, "maxAbsoluteQuantizationError": 0.030763626098632812, "zeroLogitTop8SetDifference": 7 }, { "layer": 69, "tensor": "model.layers.69.mlp.gate.e_score_correction_bias", "shard": "model-00119-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105424, "headerSha256": "144013e52a007e52e61053092b290c38fae66169b06a582e21dea32fe854e6be", "tensorByteRange": [ 1051608, 1052631 ], "tensorBytes": 1024, "tensorSha256": "6b0ebe2fd88f7bb8014e60e62afe5c7851f2e57ac22a0d6dd2ad36b1af0748b9", "rangeHttpStatus": 206, "contentRange": "bytes 1051608-1052631/5366353304", "valueCount": 256, "min": 14.644294738769531, "max": 14.96318244934082, "spread": 0.31888771057128906, "mean": 14.928789854049683, "standardDeviation": 0.03929995833271043, "fp32Distinct": 218, "bf16Distinct": 6, "bf16Min": 14.625, "bf16Max": 14.9375, "meanAbsoluteQuantizationError": 0.012290403246879578, "maxAbsoluteQuantizationError": 0.030279159545898438, "zeroLogitTop8SetDifference": 8 }, { "layer": 70, "tensor": "model.layers.70.mlp.gate.e_score_correction_bias", "shard": "model-00123-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106352, "headerSha256": "3426ff85da73d414e489b43affef3b6c60508712b440d65ca02fb04c8b674bda", "tensorByteRange": [ 634744, 635767 ], "tensorBytes": 1024, "tensorSha256": "58373779e3c585bde351d1d9b7d9bd10f90124e8af0dc2a6a54ef226f03c529c", "rangeHttpStatus": 206, "contentRange": "bytes 634744-635767/5363339256", "valueCount": 256, "min": 15.740756034851074, "max": 16.054555892944336, "spread": 0.3137998580932617, "mean": 16.016124427318573, "standardDeviation": 0.03657771582002515, "fp32Distinct": 219, "bf16Distinct": 4, "bf16Min": 15.75, "bf16Max": 16, "meanAbsoluteQuantizationError": 0.026305928826332092, "maxAbsoluteQuantizationError": 0.05455589294433594, "zeroLogitTop8SetDifference": 8 }, { "layer": 71, "tensor": "model.layers.71.mlp.gate.e_score_correction_bias", "shard": "model-00125-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105904, "headerSha256": "fa490975f4265da3fc1a333c6ac4f2d3c328fcc052eec2e83edf9f0adf71fe3c", "tensorByteRange": [ 428472, 429495 ], "tensorBytes": 1024, "tensorSha256": "caff84408ee25c80b5ea661ec0ff57e0d82ae71da8f0c7a8f5e6abededd23337", "rangeHttpStatus": 206, "contentRange": "bytes 428472-429495/5366353784", "valueCount": 256, "min": 14.555423736572266, "max": 14.83530044555664, "spread": 0.279876708984375, "mean": 14.791334543377161, "standardDeviation": 0.032815589354698464, "fp32Distinct": 228, "bf16Distinct": 5, "bf16Min": 14.5625, "bf16Max": 14.8125, "meanAbsoluteQuantizationError": 0.013332527130842209, "maxAbsoluteQuantizationError": 0.031167984008789062, "zeroLogitTop8SetDifference": 8 }, { "layer": 72, "tensor": "model.layers.72.mlp.gate.e_score_correction_bias", "shard": "model-00127-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106040, "headerSha256": "cdd6670e4c98bbf953916169d8743f3588b063ecf6fabfde6f87f46d1b3cc64c", "tensorByteRange": [ 219712, 220735 ], "tensorBytes": 1024, "tensorSha256": "823db678fd726251f5118001d91cdb54ad9ab6d89b21d8f5dee1e552572d88ce", "rangeHttpStatus": 206, "contentRange": "bytes 219712-220735/5366353920", "valueCount": 256, "min": 17.144577026367188, "max": 17.46739959716797, "spread": 0.32282257080078125, "mean": 17.423513561487198, "standardDeviation": 0.03477968861865958, "fp32Distinct": 93, "bf16Distinct": 4, "bf16Min": 17.125, "bf16Max": 17.5, "meanAbsoluteQuantizationError": 0.04506674408912659, "maxAbsoluteQuantizationError": 0.06241607666015625, "zeroLogitTop8SetDifference": 8 }, { "layer": 73, "tensor": "model.layers.73.mlp.gate.e_score_correction_bias", "shard": "model-00128-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105608, "headerSha256": "0a71c5214bfbeb179fccdf1cba3c8ccd1f582ce2f4f2c61dd28accea51237b12", "tensorByteRange": [ 1319056, 1320079 ], "tensorBytes": 1024, "tensorSha256": "6ad16edb75917cc7ec002d486a67b6cfe1af50d4b16948b4af75b4d2681d51c3", "rangeHttpStatus": 206, "contentRange": "bytes 1319056-1320079/5366353488", "valueCount": 256, "min": 19.822105407714844, "max": 20.15192413330078, "spread": 0.3298187255859375, "mean": 20.107093304395676, "standardDeviation": 0.04042584071301911, "fp32Distinct": 97, "bf16Distinct": 3, "bf16Min": 19.875, "bf16Max": 20.125, "meanAbsoluteQuantizationError": 0.019829541444778442, "maxAbsoluteQuantizationError": 0.0620269775390625, "zeroLogitTop8SetDifference": 8 }, { "layer": 74, "tensor": "model.layers.74.mlp.gate.e_score_correction_bias", "shard": "model-00130-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 106200, "headerSha256": "fbb268804254c406cebf497247b0f599fbcb2eae1d019c2a0f19be0a58f284a3", "tensorByteRange": [ 1110752, 1111775 ], "tensorBytes": 1024, "tensorSha256": "4e436b7189d0e62ca6a9193c23fe8873887ffea020640264978711a14d36ca9d", "rangeHttpStatus": 206, "contentRange": "bytes 1110752-1111775/5363339104", "valueCount": 256, "min": 13.309146881103516, "max": 13.604076385498047, "spread": 0.29492950439453125, "mean": 13.552003983408213, "standardDeviation": 0.03891983144850375, "fp32Distinct": 236, "bf16Distinct": 6, "bf16Min": 13.3125, "bf16Max": 13.625, "meanAbsoluteQuantizationError": 0.015326481312513351, "maxAbsoluteQuantizationError": 0.03104400634765625, "zeroLogitTop8SetDifference": 1 }, { "layer": 75, "tensor": "model.layers.75.mlp.gate.e_score_correction_bias", "shard": "model-00132-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105624, "headerSha256": "0936c861bfbed9310068ba06ec15dc7c7be1678aa2e31ca3af8966afd8d85812", "tensorByteRange": [ 904352, 905375 ], "tensorBytes": 1024, "tensorSha256": "8ab4d9fb1c3f7124350fed98f078bc50343580eeb8343a0091f5b3edb0d08fc0", "rangeHttpStatus": 206, "contentRange": "bytes 904352-905375/5366353504", "valueCount": 256, "min": 14.45157241821289, "max": 14.663505554199219, "spread": 0.21193313598632812, "mean": 14.60244283080101, "standardDeviation": 0.03831598433386119, "fp32Distinct": 234, "bf16Distinct": 5, "bf16Min": 14.4375, "bf16Max": 14.6875, "meanAbsoluteQuantizationError": 0.01376013457775116, "maxAbsoluteQuantizationError": 0.0310211181640625, "zeroLogitTop8SetDifference": 5 }, { "layer": 76, "tensor": "model.layers.76.mlp.gate.e_score_correction_bias", "shard": "model-00134-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105728, "headerSha256": "d58e564f85d0c841e1eb0cffa3aa57e98a80f5caffaba49a7d5040541498a519", "tensorByteRange": [ 695560, 696583 ], "tensorBytes": 1024, "tensorSha256": "ccc050baec8a627fa7263df39bcf016971b46de96b53ce98efb43f08570a5daf", "rangeHttpStatus": 206, "contentRange": "bytes 695560-696583/5366353608", "valueCount": 256, "min": 12.282735824584961, "max": 12.604593276977539, "spread": 0.3218574523925781, "mean": 12.549277991056442, "standardDeviation": 0.04345129552610395, "fp32Distinct": 238, "bf16Distinct": 5, "bf16Min": 12.3125, "bf16Max": 12.625, "meanAbsoluteQuantizationError": 0.015469737350940704, "maxAbsoluteQuantizationError": 0.031097412109375, "zeroLogitTop8SetDifference": 1 }, { "layer": 77, "tensor": "model.layers.77.mlp.gate.e_score_correction_bias", "shard": "model-00136-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 103160, "headerSha256": "d78221b25975fa6c510a2eb9aaa9d19754c83368ae7a80d2fa9fea0c4ad18926", "tensorByteRange": [ 484096, 485119 ], "tensorBytes": 1024, "tensorSha256": "bf96e8aeb9a603280b39e35e522811e07451966b05f21f335b769552e587d351", "rangeHttpStatus": 206, "contentRange": "bytes 484096-485119/5366338752", "valueCount": 256, "min": 20.146926879882812, "max": 20.59368133544922, "spread": 0.44675445556640625, "mean": 20.536615043878555, "standardDeviation": 0.06451776881492392, "fp32Distinct": 115, "bf16Distinct": 5, "bf16Min": 20.125, "bf16Max": 20.625, "meanAbsoluteQuantizationError": 0.04153159260749817, "maxAbsoluteQuantizationError": 0.0623016357421875, "zeroLogitTop8SetDifference": 7 }, { "layer": 78, "tensor": "model.layers.78.mlp.gate.e_score_correction_bias", "shard": "model-00138-of-00141.safetensors", "dtype": "F32", "shape": [ 256 ], "headerLength": 105984, "headerSha256": "62485b6f9a4fa1dda50df1456b9526b9331eb4bfdc7a5e068909d05b57b97901", "tensorByteRange": [ 314888, 315911 ], "tensorBytes": 1024, "tensorSha256": "ef9c892ebbf350e5e120c4f252e07ae6ec18e11c7e0e9d25cb8a213241483795", "rangeHttpStatus": 206, "contentRange": "bytes 314888-315911/5363351176", "valueCount": 256, "min": 15.42186164855957, "max": 15.817716598510742, "spread": 0.3958549499511719, "mean": 15.724204242229462, "standardDeviation": 0.07773809574408906, "fp32Distinct": 240, "bf16Distinct": 7, "bf16Min": 15.4375, "bf16Max": 15.8125, "meanAbsoluteQuantizationError": 0.015762202441692352, "maxAbsoluteQuantizationError": 0.031244277954101562, "zeroLogitTop8SetDifference": 8 } ], "aggregate": { "tensorCount": 76, "firstLayer": 3, "lastLayer": 78, "fp32Bytes": 77824, "bf16Bytes": 38912, "fp32MinusBf16Bytes": 38912, "fp32MinusBf16KiB": 38, "allTensorsF32": true, "allTensorsShape256": true, "allLayersLoseDistinctValuesInBf16": true, "layersWithFp32AllDistinct": 0, "medianFp32Distinct": 214, "medianBf16Distinct": 7, "minimumBf16Distinct": 3, "minimumBf16DistinctLayer": 3, "maximumBf16Distinct": 16, "maximumBf16DistinctLayer": 30, "medianSpread": 0.3203725814819336, "minimumSpread": 0.14896297454833984, "maximumSpread": 0.6426982879638672, "medianMaxAbsoluteQuantizationError": 0.03076171875, "layersWhoseZeroLogitTop8SetChanges": 75, "worstZeroLogitTop8SetDifference": 8, "downloadedWeightFraction": 1.0299395796812458e-7 } }, "syntheticFixture": { "statement": "This deterministic JavaScript fixture reimplements the numeric mechanism with synthetic linearly spaced biases and seeded random logits. It is not a GLM-5.2 checkpoint route, model call, SGLang test run, or reproduction of the PR author’s 98.50% measurement.", "expertCount": 174, "tokenCount": 64, "topK": 8, "fp32Distinct": 174, "bf16Distinct": 3, "changedTop8Sets": 60, "changedTop8Percent": 93.75 }, "deploymentGate": { "algorithm": [ "Pin the exact checkpoint, SGLang commit, AITER build, quantization configuration, architecture string, and selected top-k backend.", "Verify all e_score_correction_bias tensors remain fp32 after weight loading; checking the Hugging Face file alone is insufficient.", "Verify the selected AITER top-k call receives fp32 bias and matching fp32 gating logits; a parameter-only patch is insufficient.", "Require a merged commit in a pinned build, focused main/NextN/non-GLM tests, and a checkpoint-wide dtype assertion.", "Then compare expert IDs and output behavior under representative prompts and concurrency before bounded traffic." ], "fixtures": [ { "name": "captured-base-quantized-aiter-path", "parameterSiteFp32": false, "aiterBoundaryFp32": false, "proposalMerged": false, "releaseContainsFix": false, "checkpointAuditPassed": true, "focusedTestsPassed": false, "fullRuntimeCanaryPassed": false, "decision": "reject", "reason": "parameter-is-destroyed-before-routing" }, { "name": "parameter-site-only-hot-patch", "parameterSiteFp32": true, "aiterBoundaryFp32": false, "proposalMerged": false, "releaseContainsFix": false, "checkpointAuditPassed": true, "focusedTestsPassed": false, "fullRuntimeCanaryPassed": false, "decision": "reject", "reason": "aiter-boundary-re-downcasts-the-bias" }, { "name": "two-source-hunks-from-open-pr", "parameterSiteFp32": true, "aiterBoundaryFp32": true, "proposalMerged": false, "releaseContainsFix": false, "checkpointAuditPassed": true, "focusedTestsPassed": false, "fullRuntimeCanaryPassed": false, "decision": "hold", "reason": "source-proposal-is-open-or-not-in-a-pinned-release" }, { "name": "future-merge-without-runtime-canary", "parameterSiteFp32": true, "aiterBoundaryFp32": true, "proposalMerged": true, "releaseContainsFix": true, "checkpointAuditPassed": true, "focusedTestsPassed": true, "fullRuntimeCanaryPassed": false, "decision": "hold", "reason": "numeric-or-runtime-gate-is-incomplete" }, { "name": "future-pinned-release-and-complete-canary", "parameterSiteFp32": true, "aiterBoundaryFp32": true, "proposalMerged": true, "releaseContainsFix": true, "checkpointAuditPassed": true, "focusedTestsPassed": true, "fullRuntimeCanaryPassed": true, "decision": "eligible-for-bounded-canary", "reason": "both-dtype-sites-and-all-promotion-gates-pass" } ] }, "decision": { "currentVerdict": "The official checkpoint stores 76 routed-layer correction biases as F32, and every one loses distinct values under a deterministic BF16 conversion. SGLang PR #37133 proposes two necessary source changes but is open, unmerged, not in the pinned release, and its named aggregate workflows are red in the captured state.", "safeAction": "Keep the current production build on hold for this path, audit the loaded dtype and selected top-k backend, and promote only a merged pinned build that passes focused tests plus an end-to-end expert-ID and output canary.", "unsafeAction": "Do not hot-patch only MoEGate, do not equate a green accuracy benchmark with routing fidelity, and do not present the open PR author’s benchmark or latency numbers as independently reproduced.", "evidenceBoundary": "The audit reads public immutable metadata plus 76 exact 1,024-byte tensors, calculates numeric aggregates in Node.js, discards raw values, and does not claim runtime correctness or reproduce author benchmarks." }, "overlapAudit": { "prepublicationSitemapUrls": 99, "registryPages": 92, "distinctIntent": true, "closestPages": [ { "url": "/guides/glm-5-2-expert-parallelism/", "boundary": "owns EP sizing and communication, not correction-bias dtype" }, { "url": "/guides/glm-5-2-sglang-dsa-indexer-fusion/", "boundary": "owns DSA indexer stream safety, not MoE noaux_tc routing" }, { "url": "/guides/glm-5-2-sglang-kv-slot-zero/", "boundary": "owns MLA cache write corruption, not expert-selection bias precision" }, { "url": "/guides/glm-5-2-amd-rocm-sglang/", "boundary": "owns AMD deployment selection and provides the natural runtime backlink" } ] }, "searchSupply": { "engine": "google", "query": "\"GLM-5.2\" \"e_score_correction_bias\" SGLang", "requestId": "serpapi-e2c8e417d9df480fbf8487bbfcc0d6a7", "requests": 1, "cost": 1, "result": "failed-http", "evaluation": "failed", "changedDecision": false, "retryCount": 0 }, "aiHot": [ { "id": "cmtgihylr01tlrokdreezpex0", "permalink": "https://aihot.virxact.com/items/cmtgihylr01tlrokdreezpex0", "classification": "weak-glm-link-and-prohibited-model-development", "note": "MiniMax H3 Max live-video tooling does not create a direct GLM-5.2 routing or deployment task; MiniMax development was not authorized." }, { "id": "cmtgi3e9q01ekrokdi67kdx19", "permalink": "https://aihot.virxact.com/items/cmtgi3e9q01ekrokdi67kdx19", "classification": "duplicate-intent", "note": "The Hugging Face breach event is already covered by the site forensics guide." }, { "id": "cmtgi34hk01d3rokdi2z30urw", "permalink": "https://aihot.virxact.com/items/cmtgi34hk01d3rokdi2z30urw", "classification": "weak-glm-link", "note": "ChatGPT Work product behavior does not answer a GLM-5.2 deployment or model-selection task." } ], "discoveryChecks": { "officialZaiReleaseChecked": true, "zhipuResearchIndexChecked": true, "officialHuggingFaceCheckpointChecked": true, "upstreamSglangChecked": true }, "protectedBaseline": { "homepageSha256": "5d7f8327af1af910c0484cdd0b7fde8695a107339c207cc585452210f7fd05a8", "homepageChangeAuthorized": false }, "runtimeBoundary": { "modelCalls": 0, "fullModelShardDownloads": 0, "modelWeightBytesDownloaded": 77824, "safetensorsHeaderBytesDownloaded": 8027424, "rawTensorValuesArchived": 0, "sglangImports": 0, "torchImports": 0, "cpuInferenceRuns": 0, "gpuRuns": 0, "servingProcesses": 0, "containers": 0, "thirdPartyPatchesApplied": 0, "note": "The audit reads public immutable metadata plus 76 exact 1,024-byte tensors, calculates numeric aggregates in Node.js, discards raw values, and does not claim runtime correctness or reproduce author benchmarks." }, "invariants": { "atLeastTwentyPrimarySourceReceipts": true, "allSourcesReturned200": true, "allSourceHashesValid": true, "immutablePinsMatch": true, "proposalOpenUnmergedAndScoped": true, "authorClaimsClearlyOwned": true, "bothSourceSitesCaptured": true, "nineCpuTestsAndNextNCoverageCaptured": true, "allNamedWorkflowRunsFailedAtAggregateLevel": true, "officialModelContractPinned": true, "allSeventySixBiasTensorsRangeAudited": true, "everyLayerCollapsesUnderBf16": true, "exactWeightByteBoundary": true, "syntheticFixturePasses": true, "fiveDeploymentFixturesFailClosed": true, "distinctIntent": true, "oneFailedSerpRequestNoRetry": true, "threeAiHotItemsClassified": true, "zeroRuntimeExecution": true } }