{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PC5DMB4NHYC4LMHUY7OF7FCEIX","short_pith_number":"pith:PC5DMB4N","schema_version":"1.0","canonical_sha256":"78ba36078d3e05c5b0f4c7dc5f944445d9cf1d70acc954b8d7f257b7a98335a8","source":{"kind":"arxiv","id":"2406.14005","version":2},"attestation_state":"computed","paper":{"title":"Information Guided Regularization for Fine-tuning Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Mandar Sharma, Naren Ramakrishnan, Nikhil Muralidhar, Raquib Bin Yousuf, Shengzhe Xu","submitted_at":"2024-06-20T05:18:37Z","abstract_excerpt":"The pretraining-fine-tuning paradigm has been the de facto strategy for transfer learning in modern language modeling. With the understanding that task adaptation in LMs is often a function of parameters shared across tasks, we argue that a more surgical approach to regularization needs to exist for smoother transfer learning. Towards this end, we investigate how the pretraining loss landscape is affected by these task-sensitive parameters through an information-theoretic lens. We then leverage the findings from our investigations to devise a novel approach to dropout for improved model regula"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2406.14005","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-20T05:18:37Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"020cd00ad47b828a131c40d5cbd56ca1c4ad070034d339a716f4903a7d42a3e4","abstract_canon_sha256":"331cbb7ec41e7d337d76e3e948d750cbebfdc8b458a62a1f7857e0c67336867f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:35:04.834690Z","signature_b64":"s06L2V82dn5yT9hpBTcq2myzMV5CoWB3fmUAt/UC4pyPNRVGMTcrn/nnMstNEam8mN30wnq3JdVajyMk9etWCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"78ba36078d3e05c5b0f4c7dc5f944445d9cf1d70acc954b8d7f257b7a98335a8","last_reissued_at":"2026-07-05T08:35:04.834187Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:35:04.834187Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Information Guided Regularization for Fine-tuning Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Mandar Sharma, Naren Ramakrishnan, Nikhil Muralidhar, Raquib Bin Yousuf, Shengzhe Xu","submitted_at":"2024-06-20T05:18:37Z","abstract_excerpt":"The pretraining-fine-tuning paradigm has been the de facto strategy for transfer learning in modern language modeling. With the understanding that task adaptation in LMs is often a function of parameters shared across tasks, we argue that a more surgical approach to regularization needs to exist for smoother transfer learning. Towards this end, we investigate how the pretraining loss landscape is affected by these task-sensitive parameters through an information-theoretic lens. We then leverage the findings from our investigations to devise a novel approach to dropout for improved model regula"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.14005","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2406.14005/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2406.14005","created_at":"2026-07-05T08:35:04.834243+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.14005v2","created_at":"2026-07-05T08:35:04.834243+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.14005","created_at":"2026-07-05T08:35:04.834243+00:00"},{"alias_kind":"pith_short_12","alias_value":"PC5DMB4NHYC4","created_at":"2026-07-05T08:35:04.834243+00:00"},{"alias_kind":"pith_short_16","alias_value":"PC5DMB4NHYC4LMHU","created_at":"2026-07-05T08:35:04.834243+00:00"},{"alias_kind":"pith_short_8","alias_value":"PC5DMB4N","created_at":"2026-07-05T08:35:04.834243+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.13262","citing_title":"Chem-GMNet: A Sphere-Native Geometric Transformer for Molecular Property Prediction","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX","json":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX.json","graph_json":"https://pith.science/api/pith-number/PC5DMB4NHYC4LMHUY7OF7FCEIX/graph.json","events_json":"https://pith.science/api/pith-number/PC5DMB4NHYC4LMHUY7OF7FCEIX/events.json","paper":"https://pith.science/paper/PC5DMB4N"},"agent_actions":{"view_html":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX","download_json":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX.json","view_paper":"https://pith.science/paper/PC5DMB4N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.14005&json=true","fetch_graph":"https://pith.science/api/pith-number/PC5DMB4NHYC4LMHUY7OF7FCEIX/graph.json","fetch_events":"https://pith.science/api/pith-number/PC5DMB4NHYC4LMHUY7OF7FCEIX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/action/storage_attestation","attest_author":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/action/author_attestation","sign_citation":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/action/citation_signature","submit_replication":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/action/replication_record"}},"created_at":"2026-07-05T08:35:04.834243+00:00","updated_at":"2026-07-05T08:35:04.834243+00:00"}