{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:7YTDWIM7RKAEGPLVOA5TXKCDQK","short_pith_number":"pith:7YTDWIM7","canonical_record":{"source":{"id":"2502.02013","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-04T05:03:42Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"9a78a8101cc4fdc413f47784f866b9506c7f104983fbe8b6d18d8b7782bb0377","abstract_canon_sha256":"a12f5f2a407f00d1293ae8d608f79d7f44ea909c8948eb34baa5bdaf7cb40d41"},"schema_version":"1.0"},"canonical_sha256":"fe263b219f8a80433d75703b3ba84382a8ad6d6c910f2ab1c3801b05f4583f8f","source":{"kind":"arxiv","id":"2502.02013","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.02013","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"arxiv_version","alias_value":"2502.02013v2","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.02013","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"pith_short_12","alias_value":"7YTDWIM7RKAE","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"7YTDWIM7RKAEGPLV","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"7YTDWIM7","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:7YTDWIM7RKAEGPLVOA5TXKCDQK","target":"record","payload":{"canonical_record":{"source":{"id":"2502.02013","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-04T05:03:42Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"9a78a8101cc4fdc413f47784f866b9506c7f104983fbe8b6d18d8b7782bb0377","abstract_canon_sha256":"a12f5f2a407f00d1293ae8d608f79d7f44ea909c8948eb34baa5bdaf7cb40d41"},"schema_version":"1.0"},"canonical_sha256":"fe263b219f8a80433d75703b3ba84382a8ad6d6c910f2ab1c3801b05f4583f8f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:50.917031Z","signature_b64":"DIMIDjDO4dheiJKFR84HIuHEkB/PuuOMjrKWD0DxvVfb2qGx296LqBmXlgaKaCyhieEHGE/kOIFTIoOZs/R1BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fe263b219f8a80433d75703b3ba84382a8ad6d6c910f2ab1c3801b05f4583f8f","last_reissued_at":"2026-05-17T23:38:50.916364Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:50.916364Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.02013","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ghrNxHOoN/h9NyK3ToSq05gDrE7eSNv7cEsZdgAsMNjuSHnLSjsLzul/F21DfIkS9QQ+D325yWJ36f5APr90CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T13:22:47.051846Z"},"content_sha256":"20eda92a8e67d25c04a72a4ac3dc61359bd369154efef8f6d51dd2718bc49014","schema_version":"1.0","event_id":"sha256:20eda92a8e67d25c04a72a4ac3dc61359bd369154efef8f6d51dd2718bc49014"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:7YTDWIM7RKAEGPLVOA5TXKCDQK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Layer by Layer: Uncovering Hidden Representations in Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"Intermediate layers in language models often encode richer representations than the final layer for downstream tasks.","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Dan Zhao, Jalal Naghiyev, Md Rifat Arefin, Niket Patel, Oscar Skean, Ravid Shwartz-Ziv, Yann LeCun","submitted_at":"2025-02-04T05:03:42Z","abstract_excerpt":"From extracting features to generating text, the outputs of large language models (LLMs) typically rely on the final layers, following the conventional wisdom that earlier layers capture only low-level cues. However, our analysis shows that intermediate layers can encode even richer representations, often improving performance on a range of downstream tasks. To explain and quantify these hidden-layer properties, we propose a unified framework of representation quality metrics based on information theory, geometry, and invariance to input perturbations. Our framework highlights how each layer b"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"our analysis shows that intermediate layers can encode even richer representations, often improving performance on a range of downstream tasks... intermediate layers consistently provide stronger features, challenging the standard view on final-layer embeddings","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"that the proposed metrics based on information theory, geometry, and invariance to input perturbations accurately capture representation quality relevant to downstream task performance","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Intermediate layers in LLMs consistently provide stronger features than final layers across tasks and architectures, as quantified by a new framework of information-theoretic, geometric, and invariance metrics.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Intermediate layers in language models often encode richer representations than the final layer for downstream tasks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"dc8b59d945f23b502183cf63228d93b678d15de36e79b0be330c750dd4d9f48f"},"source":{"id":"2502.02013","kind":"arxiv","version":2},"verdict":{"id":"e9508d18-ead8-4936-a86f-0f4c9811f422","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T16:25:43.473004Z","strongest_claim":"our analysis shows that intermediate layers can encode even richer representations, often improving performance on a range of downstream tasks... intermediate layers consistently provide stronger features, challenging the standard view on final-layer embeddings","one_line_summary":"Intermediate layers in LLMs consistently provide stronger features than final layers across tasks and architectures, as quantified by a new framework of information-theoretic, geometric, and invariance metrics.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"that the proposed metrics based on information theory, geometry, and invariance to input perturbations accurately capture representation quality relevant to downstream task performance","pith_extraction_headline":"Intermediate layers in language models often encode richer representations than the final layer for downstream tasks."},"references":{"count":173,"sample":[{"doi":"","year":2022,"title":"Agrawal, K. K., Mondal, A. K., Ghosh, A., and Richards, B. - ReQ : Assessing representation quality in self-supervised learning by measuring eigenspectrum decay. NeurIPs, 2022","work_id":"8fe0fb88-86eb-414b-b26d-362bdcf0cf9e","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2017,"title":"Alain, G. and Bengio, Y. Understanding intermediate layers using linear classifier probes. ICLR, 2017","work_id":"a916bb7d-1cf2-481f-88c9-5d8cc9240c10","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"R., Subbaraj, G., Gontier, N., LeCun, Y., Rish, I., Shwartz-Ziv, R., and Pal, C","work_id":"b4a8c8bf-2a6e-433d-8438-d1898900a6ac","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Information theory with kernel methods","work_id":"cd38bb25-b39e-4543-8a17-5c6e2db8d88b","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"BeIT : Bert pre-training of image transformers","work_id":"c29e2625-5fb6-47c4-b15e-ad4380212f00","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":173,"snapshot_sha256":"4fc9c8e093cc20e5ad7db16f3600410d3e100fea8d14d137fc6435efbe234874","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"52234e52e1e85226baf11be515a74ec46aff0d131f1979ad36a5a9b8285dd307"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"e9508d18-ead8-4936-a86f-0f4c9811f422"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"AP3JmoBZxd3J8miu2aS/muPIz3knniiLRpQ7jUPmOjgVP3KJiiOCXn+SrZKBohLhbtYYblyFht30XyAk3QlnBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T13:22:47.052604Z"},"content_sha256":"1eba8a99c990f83a79d4223ad40dce6a4c136e776f2f31b638c9daa0d37549aa","schema_version":"1.0","event_id":"sha256:1eba8a99c990f83a79d4223ad40dce6a4c136e776f2f31b638c9daa0d37549aa"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/7YTDWIM7RKAEGPLVOA5TXKCDQK/bundle.json","state_url":"https://pith.science/pith/7YTDWIM7RKAEGPLVOA5TXKCDQK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/7YTDWIM7RKAEGPLVOA5TXKCDQK/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-25T13:22:47Z","links":{"resolver":"https://pith.science/pith/7YTDWIM7RKAEGPLVOA5TXKCDQK","bundle":"https://pith.science/pith/7YTDWIM7RKAEGPLVOA5TXKCDQK/bundle.json","state":"https://pith.science/pith/7YTDWIM7RKAEGPLVOA5TXKCDQK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/7YTDWIM7RKAEGPLVOA5TXKCDQK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:7YTDWIM7RKAEGPLVOA5TXKCDQK","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"a12f5f2a407f00d1293ae8d608f79d7f44ea909c8948eb34baa5bdaf7cb40d41","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-04T05:03:42Z","title_canon_sha256":"9a78a8101cc4fdc413f47784f866b9506c7f104983fbe8b6d18d8b7782bb0377"},"schema_version":"1.0","source":{"id":"2502.02013","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.02013","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"arxiv_version","alias_value":"2502.02013v2","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.02013","created_at":"2026-05-17T23:38:50Z"},{"alias_kind":"pith_short_12","alias_value":"7YTDWIM7RKAE","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"7YTDWIM7RKAEGPLV","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"7YTDWIM7","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:1eba8a99c990f83a79d4223ad40dce6a4c136e776f2f31b638c9daa0d37549aa","target":"graph","created_at":"2026-05-17T23:38:50Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"our analysis shows that intermediate layers can encode even richer representations, often improving performance on a range of downstream tasks... intermediate layers consistently provide stronger features, challenging the standard view on final-layer embeddings"},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"that the proposed metrics based on information theory, geometry, and invariance to input perturbations accurately capture representation quality relevant to downstream task performance"},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Intermediate layers in LLMs consistently provide stronger features than final layers across tasks and architectures, as quantified by a new framework of information-theoretic, geometric, and invariance metrics."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Intermediate layers in language models often encode richer representations than the final layer for downstream tasks."}],"snapshot_sha256":"dc8b59d945f23b502183cf63228d93b678d15de36e79b0be330c750dd4d9f48f"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"52234e52e1e85226baf11be515a74ec46aff0d131f1979ad36a5a9b8285dd307"},"paper":{"abstract_excerpt":"From extracting features to generating text, the outputs of large language models (LLMs) typically rely on the final layers, following the conventional wisdom that earlier layers capture only low-level cues. However, our analysis shows that intermediate layers can encode even richer representations, often improving performance on a range of downstream tasks. To explain and quantify these hidden-layer properties, we propose a unified framework of representation quality metrics based on information theory, geometry, and invariance to input perturbations. Our framework highlights how each layer b","authors_text":"Dan Zhao, Jalal Naghiyev, Md Rifat Arefin, Niket Patel, Oscar Skean, Ravid Shwartz-Ziv, Yann LeCun","cross_cats":["cs.AI","cs.CL"],"headline":"Intermediate layers in language models often encode richer representations than the final layer for downstream tasks.","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-04T05:03:42Z","title":"Layer by Layer: Uncovering Hidden Representations in Language Models"},"references":{"count":173,"internal_anchors":0,"resolved_work":173,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Agrawal, K. K., Mondal, A. K., Ghosh, A., and Richards, B. - ReQ : Assessing representation quality in self-supervised learning by measuring eigenspectrum decay. NeurIPs, 2022","work_id":"8fe0fb88-86eb-414b-b26d-362bdcf0cf9e","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Alain, G. and Bengio, Y. Understanding intermediate layers using linear classifier probes. ICLR, 2017","work_id":"a916bb7d-1cf2-481f-88c9-5d8cc9240c10","year":2017},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"R., Subbaraj, G., Gontier, N., LeCun, Y., Rish, I., Shwartz-Ziv, R., and Pal, C","work_id":"b4a8c8bf-2a6e-433d-8438-d1898900a6ac","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Information theory with kernel methods","work_id":"cd38bb25-b39e-4543-8a17-5c6e2db8d88b","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"BeIT : Bert pre-training of image transformers","work_id":"c29e2625-5fb6-47c4-b15e-ad4380212f00","year":2022}],"snapshot_sha256":"4fc9c8e093cc20e5ad7db16f3600410d3e100fea8d14d137fc6435efbe234874"},"source":{"id":"2502.02013","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-15T16:25:43.473004Z","id":"e9508d18-ead8-4936-a86f-0f4c9811f422","model_set":{"reader":"grok-4.3"},"one_line_summary":"Intermediate layers in LLMs consistently provide stronger features than final layers across tasks and architectures, as quantified by a new framework of information-theoretic, geometric, and invariance metrics.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Intermediate layers in language models often encode richer representations than the final layer for downstream tasks.","strongest_claim":"our analysis shows that intermediate layers can encode even richer representations, often improving performance on a range of downstream tasks... intermediate layers consistently provide stronger features, challenging the standard view on final-layer embeddings","weakest_assumption":"that the proposed metrics based on information theory, geometry, and invariance to input perturbations accurately capture representation quality relevant to downstream task performance"}},"verdict_id":"e9508d18-ead8-4936-a86f-0f4c9811f422"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:20eda92a8e67d25c04a72a4ac3dc61359bd369154efef8f6d51dd2718bc49014","target":"record","created_at":"2026-05-17T23:38:50Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"a12f5f2a407f00d1293ae8d608f79d7f44ea909c8948eb34baa5bdaf7cb40d41","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-04T05:03:42Z","title_canon_sha256":"9a78a8101cc4fdc413f47784f866b9506c7f104983fbe8b6d18d8b7782bb0377"},"schema_version":"1.0","source":{"id":"2502.02013","kind":"arxiv","version":2}},"canonical_sha256":"fe263b219f8a80433d75703b3ba84382a8ad6d6c910f2ab1c3801b05f4583f8f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"fe263b219f8a80433d75703b3ba84382a8ad6d6c910f2ab1c3801b05f4583f8f","first_computed_at":"2026-05-17T23:38:50.916364Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:50.916364Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"DIMIDjDO4dheiJKFR84HIuHEkB/PuuOMjrKWD0DxvVfb2qGx296LqBmXlgaKaCyhieEHGE/kOIFTIoOZs/R1BA==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:50.917031Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.02013","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:20eda92a8e67d25c04a72a4ac3dc61359bd369154efef8f6d51dd2718bc49014","sha256:1eba8a99c990f83a79d4223ad40dce6a4c136e776f2f31b638c9daa0d37549aa"],"state_sha256":"272025d9d10ec61464ce7818b5a0df217656f8ca7ef132680be40fdaa45bbd04"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YqY11N/RVstsiEJ9V4Yyty54c1xrqrjj3sKzmohDJfbDv2BbLU0//CB521z0llXoSuNcGR2Mi2B3gyTxvjY1Bg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-25T13:22:47.057334Z","bundle_sha256":"937005bc9989f1153208b86321ef0a2f11fa70df0e9db401d854da778b44dbb0"}}