{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:PC5DMB4NHYC4LMHUY7OF7FCEIX","short_pith_number":"pith:PC5DMB4N","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"},"canonical_sha256":"78ba36078d3e05c5b0f4c7dc5f944445d9cf1d70acc954b8d7f257b7a98335a8","source":{"kind":"arxiv","id":"2406.14005","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.14005","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"arxiv_version","alias_value":"2406.14005v2","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.14005","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"pith_short_12","alias_value":"PC5DMB4NHYC4","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"pith_short_16","alias_value":"PC5DMB4NHYC4LMHU","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"pith_short_8","alias_value":"PC5DMB4N","created_at":"2026-07-05T08:35:04Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:PC5DMB4NHYC4LMHUY7OF7FCEIX","target":"record","payload":{"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"},"canonical_sha256":"78ba36078d3e05c5b0f4c7dc5f944445d9cf1d70acc954b8d7f257b7a98335a8","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"},"source_kind":"arxiv","source_id":"2406.14005","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":"2026-07-05T08:35:04Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"EHFbNVgD3D5CA/p74RkLGZMz6jVOezolrtKI03CELvf0VgiOC7qnDLnz71wB6ABG7+lgeA2SEFKxNvG5yGyHCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-09T02:37:02.381417Z"},"content_sha256":"a2b5f68b15146709f434e9fa3b06bdc4f632c85b18b7186ed5851663444eefbf","schema_version":"1.0","event_id":"sha256:a2b5f68b15146709f434e9fa3b06bdc4f632c85b18b7186ed5851663444eefbf"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:PC5DMB4NHYC4LMHUY7OF7FCEIX","target":"graph","payload":{"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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:35:04Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"e52sxbblHfaU9Tnk4ERbeojvzyFKXpk+hvbbWi+3SF1/xkMLoM53Ca6R3xZc1rgl9U5Bwmwat47hR+VQs7jmBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-09T02:37:02.381804Z"},"content_sha256":"252935c4f85ba788876d1c45c38b506e1323dcbb8c9637cec91a770354f4c81c","schema_version":"1.0","event_id":"sha256:252935c4f85ba788876d1c45c38b506e1323dcbb8c9637cec91a770354f4c81c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/bundle.json","state_url":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/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":"2026-07-09T02:37:02Z","links":{"resolver":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX","bundle":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/bundle.json","state":"https://pith.science/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/PC5DMB4NHYC4LMHUY7OF7FCEIX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:PC5DMB4NHYC4LMHUY7OF7FCEIX","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":"331cbb7ec41e7d337d76e3e948d750cbebfdc8b458a62a1f7857e0c67336867f","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-20T05:18:37Z","title_canon_sha256":"020cd00ad47b828a131c40d5cbd56ca1c4ad070034d339a716f4903a7d42a3e4"},"schema_version":"1.0","source":{"id":"2406.14005","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.14005","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"arxiv_version","alias_value":"2406.14005v2","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.14005","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"pith_short_12","alias_value":"PC5DMB4NHYC4","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"pith_short_16","alias_value":"PC5DMB4NHYC4LMHU","created_at":"2026-07-05T08:35:04Z"},{"alias_kind":"pith_short_8","alias_value":"PC5DMB4N","created_at":"2026-07-05T08:35:04Z"}],"graph_snapshots":[{"event_id":"sha256:252935c4f85ba788876d1c45c38b506e1323dcbb8c9637cec91a770354f4c81c","target":"graph","created_at":"2026-07-05T08:35:04Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2406.14005/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"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","authors_text":"Mandar Sharma, Naren Ramakrishnan, Nikhil Muralidhar, Raquib Bin Yousuf, Shengzhe Xu","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-20T05:18:37Z","title":"Information Guided Regularization for Fine-tuning Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.14005","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:a2b5f68b15146709f434e9fa3b06bdc4f632c85b18b7186ed5851663444eefbf","target":"record","created_at":"2026-07-05T08:35:04Z","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":"331cbb7ec41e7d337d76e3e948d750cbebfdc8b458a62a1f7857e0c67336867f","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-20T05:18:37Z","title_canon_sha256":"020cd00ad47b828a131c40d5cbd56ca1c4ad070034d339a716f4903a7d42a3e4"},"schema_version":"1.0","source":{"id":"2406.14005","kind":"arxiv","version":2}},"canonical_sha256":"78ba36078d3e05c5b0f4c7dc5f944445d9cf1d70acc954b8d7f257b7a98335a8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"78ba36078d3e05c5b0f4c7dc5f944445d9cf1d70acc954b8d7f257b7a98335a8","first_computed_at":"2026-07-05T08:35:04.834187Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:35:04.834187Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"s06L2V82dn5yT9hpBTcq2myzMV5CoWB3fmUAt/UC4pyPNRVGMTcrn/nnMstNEam8mN30wnq3JdVajyMk9etWCA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:35:04.834690Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.14005","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a2b5f68b15146709f434e9fa3b06bdc4f632c85b18b7186ed5851663444eefbf","sha256:252935c4f85ba788876d1c45c38b506e1323dcbb8c9637cec91a770354f4c81c"],"state_sha256":"2308dbfc1a9daada20820781ed0e9cf3d3a15ec89b5560d3142e74fe7193430b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"raj7/CoJZ4jMkBa7EO6rX3a5GM1p+5xzbLzVQ+nINhdAORTlJqHENCqQ2tJ/p7zGnP0fEAGSUxEQtekgVTQUDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-09T02:37:02.383873Z","bundle_sha256":"d0a542091915f564d2cb7384c2d4d1d5be74b5d9796697c3125da95e2c8d6429"}}