{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:P62HU4TDSBCENIHV3I2TV7JKJU","short_pith_number":"pith:P62HU4TD","canonical_record":{"source":{"id":"2308.13320","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-08-25T11:49:51Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"a37cc7e790f845e8b28236b49ccb1da0a562c7d3691ce0bd3dc822b822887c3e","abstract_canon_sha256":"c8d6ec8a23aee71add7dc33a1a19d1bc9862b98be3893bc90d760cc8c0cae0a3"},"schema_version":"1.0"},"canonical_sha256":"7fb47a7263904446a0f5da353afd2a4d0d4638bbf1894592ab8886fb368a84f3","source":{"kind":"arxiv","id":"2308.13320","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2308.13320","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"arxiv_version","alias_value":"2308.13320v3","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2308.13320","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"pith_short_12","alias_value":"P62HU4TDSBCE","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"pith_short_16","alias_value":"P62HU4TDSBCENIHV","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"pith_short_8","alias_value":"P62HU4TD","created_at":"2026-07-05T08:38:34Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:P62HU4TDSBCENIHV3I2TV7JKJU","target":"record","payload":{"canonical_record":{"source":{"id":"2308.13320","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-08-25T11:49:51Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"a37cc7e790f845e8b28236b49ccb1da0a562c7d3691ce0bd3dc822b822887c3e","abstract_canon_sha256":"c8d6ec8a23aee71add7dc33a1a19d1bc9862b98be3893bc90d760cc8c0cae0a3"},"schema_version":"1.0"},"canonical_sha256":"7fb47a7263904446a0f5da353afd2a4d0d4638bbf1894592ab8886fb368a84f3","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:38:34.138074Z","signature_b64":"PWzvmLZR8a3anG14xGG4Lccft0ym+ZtJp9fd4bkfxQVvMqQFOLoSAeqzVuDh/WZjucD7YU3wW/wuVIxZFWYlDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7fb47a7263904446a0f5da353afd2a4d0d4638bbf1894592ab8886fb368a84f3","last_reissued_at":"2026-07-05T08:38:34.137542Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:38:34.137542Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2308.13320","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:38:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"sEtadkATKa1bgjGBWgwBzpAycuIQ9XQDIiougvd0iLyldqEjfyVALSsmLbuIgny6hM0YpmTy7wjwXztITYuKDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T01:49:25.164698Z"},"content_sha256":"b8aba444526548185fe7a83fcdc5d2e02d1036bc352413df428112cdb06b8ef1","schema_version":"1.0","event_id":"sha256:b8aba444526548185fe7a83fcdc5d2e02d1036bc352413df428112cdb06b8ef1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:P62HU4TDSBCENIHV3I2TV7JKJU","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Fine-tuning can cripple your foundation model; preserving features may be the solution","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Jishnu Mukhoti, Philip H.S. Torr, Puneet K. Dokania, Yarin Gal","submitted_at":"2023-08-25T11:49:51Z","abstract_excerpt":"Pre-trained foundation models, due to their enormous capacity and exposure to vast amounts of data during pre-training, are known to have learned plenty of real-world concepts. An important step in making these pre-trained models effective on downstream tasks is to fine-tune them on related datasets. While various fine-tuning methods have been devised and have been shown to be highly effective, we observe that a fine-tuned model's ability to recognize concepts on tasks $\\textit{different}$ from the downstream one is reduced significantly compared to its pre-trained counterpart. This is an unde"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2308.13320","kind":"arxiv","version":3},"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/2308.13320/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":"2026-07-05T08:38:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cp9x/lVOzBMmzZKR2voD6OuCn+RVwVajBJsWltbYaqJ5quGfD4t2KNy5N/NsZUrtoaAiDCP7XtbkOlx07SRYCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T01:49:25.165178Z"},"content_sha256":"db56dedca48886244b7f728a6aacd60d59c86b6587a6ae15e288db715ea90722","schema_version":"1.0","event_id":"sha256:db56dedca48886244b7f728a6aacd60d59c86b6587a6ae15e288db715ea90722"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/P62HU4TDSBCENIHV3I2TV7JKJU/bundle.json","state_url":"https://pith.science/pith/P62HU4TDSBCENIHV3I2TV7JKJU/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/P62HU4TDSBCENIHV3I2TV7JKJU/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":"2026-08-10T01:49:25Z","links":{"resolver":"https://pith.science/pith/P62HU4TDSBCENIHV3I2TV7JKJU","bundle":"https://pith.science/pith/P62HU4TDSBCENIHV3I2TV7JKJU/bundle.json","state":"https://pith.science/pith/P62HU4TDSBCENIHV3I2TV7JKJU/state.json","well_known_bundle":"https://pith.science/.well-known/pith/P62HU4TDSBCENIHV3I2TV7JKJU/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:P62HU4TDSBCENIHV3I2TV7JKJU","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":"c8d6ec8a23aee71add7dc33a1a19d1bc9862b98be3893bc90d760cc8c0cae0a3","cross_cats_sorted":["cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-08-25T11:49:51Z","title_canon_sha256":"a37cc7e790f845e8b28236b49ccb1da0a562c7d3691ce0bd3dc822b822887c3e"},"schema_version":"1.0","source":{"id":"2308.13320","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2308.13320","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"arxiv_version","alias_value":"2308.13320v3","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2308.13320","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"pith_short_12","alias_value":"P62HU4TDSBCE","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"pith_short_16","alias_value":"P62HU4TDSBCENIHV","created_at":"2026-07-05T08:38:34Z"},{"alias_kind":"pith_short_8","alias_value":"P62HU4TD","created_at":"2026-07-05T08:38:34Z"}],"graph_snapshots":[{"event_id":"sha256:db56dedca48886244b7f728a6aacd60d59c86b6587a6ae15e288db715ea90722","target":"graph","created_at":"2026-07-05T08:38:34Z","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/2308.13320/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pre-trained foundation models, due to their enormous capacity and exposure to vast amounts of data during pre-training, are known to have learned plenty of real-world concepts. An important step in making these pre-trained models effective on downstream tasks is to fine-tune them on related datasets. While various fine-tuning methods have been devised and have been shown to be highly effective, we observe that a fine-tuned model's ability to recognize concepts on tasks $\\textit{different}$ from the downstream one is reduced significantly compared to its pre-trained counterpart. This is an unde","authors_text":"Jishnu Mukhoti, Philip H.S. Torr, Puneet K. Dokania, Yarin Gal","cross_cats":["cs.CV"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-08-25T11:49:51Z","title":"Fine-tuning can cripple your foundation model; preserving features may be the solution"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2308.13320","kind":"arxiv","version":3},"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:b8aba444526548185fe7a83fcdc5d2e02d1036bc352413df428112cdb06b8ef1","target":"record","created_at":"2026-07-05T08:38:34Z","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":"c8d6ec8a23aee71add7dc33a1a19d1bc9862b98be3893bc90d760cc8c0cae0a3","cross_cats_sorted":["cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-08-25T11:49:51Z","title_canon_sha256":"a37cc7e790f845e8b28236b49ccb1da0a562c7d3691ce0bd3dc822b822887c3e"},"schema_version":"1.0","source":{"id":"2308.13320","kind":"arxiv","version":3}},"canonical_sha256":"7fb47a7263904446a0f5da353afd2a4d0d4638bbf1894592ab8886fb368a84f3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7fb47a7263904446a0f5da353afd2a4d0d4638bbf1894592ab8886fb368a84f3","first_computed_at":"2026-07-05T08:38:34.137542Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:38:34.137542Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PWzvmLZR8a3anG14xGG4Lccft0ym+ZtJp9fd4bkfxQVvMqQFOLoSAeqzVuDh/WZjucD7YU3wW/wuVIxZFWYlDg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:38:34.138074Z","signed_message":"canonical_sha256_bytes"},"source_id":"2308.13320","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b8aba444526548185fe7a83fcdc5d2e02d1036bc352413df428112cdb06b8ef1","sha256:db56dedca48886244b7f728a6aacd60d59c86b6587a6ae15e288db715ea90722"],"state_sha256":"944b3ad6622f7454248be2328d2c79348572870539608011e4c90d8bbba29ee8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MpUkg1M2TjZ2ApKJs0Uv5aLpiI1wHS6wmvFYtf+2UjCBzmCSqSQWxjBJQyWrxuqS9ZfZkDLg178qMWH65xRkCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T01:49:25.168518Z","bundle_sha256":"94b58160db1f2d69b592a4d3b09d17526a95c63351de257827674692300798fa"}}