{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:DSO37KUNKAOIOAINV6KEV75IA4","short_pith_number":"pith:DSO37KUN","canonical_record":{"source":{"id":"2502.02797","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-05T00:49:59Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"f2024857868a8db2102729c8a6a2a2645776bbc9f984e8cde1aa30be4a98965c","abstract_canon_sha256":"d4d85db667ced0e2c54af06eb0d9628ce27d76d265799e63cddbb7a2b96a67f4"},"schema_version":"1.0"},"canonical_sha256":"1c9dbfaa8d501c87010daf944affa807311ce5018cae046a1c39c1166b6742f8","source":{"kind":"arxiv","id":"2502.02797","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.02797","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"arxiv_version","alias_value":"2502.02797v2","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.02797","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"pith_short_12","alias_value":"DSO37KUNKAOI","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"pith_short_16","alias_value":"DSO37KUNKAOIOAIN","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"pith_short_8","alias_value":"DSO37KUN","created_at":"2026-07-05T11:20:08Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:DSO37KUNKAOIOAINV6KEV75IA4","target":"record","payload":{"canonical_record":{"source":{"id":"2502.02797","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-05T00:49:59Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"f2024857868a8db2102729c8a6a2a2645776bbc9f984e8cde1aa30be4a98965c","abstract_canon_sha256":"d4d85db667ced0e2c54af06eb0d9628ce27d76d265799e63cddbb7a2b96a67f4"},"schema_version":"1.0"},"canonical_sha256":"1c9dbfaa8d501c87010daf944affa807311ce5018cae046a1c39c1166b6742f8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:20:08.730958Z","signature_b64":"Sra590f9nXzO8YT7myZwXplo80mVr8MGHwBKCMliJSj5/tZtaC8XJhDDfDXeQW+ZBrUutVPtz30kTBDAJEEqAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1c9dbfaa8d501c87010daf944affa807311ce5018cae046a1c39c1166b6742f8","last_reissued_at":"2026-07-05T11:20:08.730485Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:20:08.730485Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.02797","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":"2026-07-05T11:20:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"w5aTklDZKJzdpoGfw+Nrbz6IQx8+SitH4gCe8FDuX/JNEY4+e+6rWvc/zJfYSZtR0e+CEoWH2A8AI04sL1mwAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T03:35:27.336761Z"},"content_sha256":"0d0fbf2ea4b055a0d2608ba82a7216d806bde13ba1229c5f1a97f9f352e1e49d","schema_version":"1.0","event_id":"sha256:0d0fbf2ea4b055a0d2608ba82a7216d806bde13ba1229c5f1a97f9f352e1e49d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:DSO37KUNKAOIOAINV6KEV75IA4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Upweighting Easy Samples in Fine-Tuning Mitigates Forgetting","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ali Kavis, Hayden Prairie, Rudrajit Das, Sujay Sanghavi, Sunny Sanyal","submitted_at":"2025-02-05T00:49:59Z","abstract_excerpt":"Fine-tuning a pre-trained model on a downstream task often degrades its original capabilities, a phenomenon known as \"catastrophic forgetting\". This is especially an issue when one does not have access to the data and recipe used to develop the pre-trained model. Under this constraint, most existing methods for mitigating forgetting are inapplicable. To address this challenge, we propose a sample weighting scheme for the fine-tuning data solely based on the pre-trained model's losses. Specifically, we upweight the easy samples on which the pre-trained model's loss is low and vice versa to limi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.02797","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/2502.02797/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":"2026-07-05T11:20:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"M7F5n1Q+GnfvJ8DVDgCq/Q4r2r4nexMp0r7ExrY9tX3kt2fhi3+VtNupyixgwYPqa2g+bqEN9c+l0utcvgCIDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T03:35:27.337616Z"},"content_sha256":"a863776cbc56e4b4cec30b7df821bc9903bf019fa184133d504be3fa9ca3cdc2","schema_version":"1.0","event_id":"sha256:a863776cbc56e4b4cec30b7df821bc9903bf019fa184133d504be3fa9ca3cdc2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DSO37KUNKAOIOAINV6KEV75IA4/bundle.json","state_url":"https://pith.science/pith/DSO37KUNKAOIOAINV6KEV75IA4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DSO37KUNKAOIOAINV6KEV75IA4/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":"2026-08-10T03:35:27Z","links":{"resolver":"https://pith.science/pith/DSO37KUNKAOIOAINV6KEV75IA4","bundle":"https://pith.science/pith/DSO37KUNKAOIOAINV6KEV75IA4/bundle.json","state":"https://pith.science/pith/DSO37KUNKAOIOAINV6KEV75IA4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DSO37KUNKAOIOAINV6KEV75IA4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:DSO37KUNKAOIOAINV6KEV75IA4","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":"d4d85db667ced0e2c54af06eb0d9628ce27d76d265799e63cddbb7a2b96a67f4","cross_cats_sorted":["cs.AI","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-05T00:49:59Z","title_canon_sha256":"f2024857868a8db2102729c8a6a2a2645776bbc9f984e8cde1aa30be4a98965c"},"schema_version":"1.0","source":{"id":"2502.02797","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.02797","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"arxiv_version","alias_value":"2502.02797v2","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.02797","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"pith_short_12","alias_value":"DSO37KUNKAOI","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"pith_short_16","alias_value":"DSO37KUNKAOIOAIN","created_at":"2026-07-05T11:20:08Z"},{"alias_kind":"pith_short_8","alias_value":"DSO37KUN","created_at":"2026-07-05T11:20:08Z"}],"graph_snapshots":[{"event_id":"sha256:a863776cbc56e4b4cec30b7df821bc9903bf019fa184133d504be3fa9ca3cdc2","target":"graph","created_at":"2026-07-05T11:20:08Z","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/2502.02797/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Fine-tuning a pre-trained model on a downstream task often degrades its original capabilities, a phenomenon known as \"catastrophic forgetting\". This is especially an issue when one does not have access to the data and recipe used to develop the pre-trained model. Under this constraint, most existing methods for mitigating forgetting are inapplicable. To address this challenge, we propose a sample weighting scheme for the fine-tuning data solely based on the pre-trained model's losses. Specifically, we upweight the easy samples on which the pre-trained model's loss is low and vice versa to limi","authors_text":"Ali Kavis, Hayden Prairie, Rudrajit Das, Sujay Sanghavi, Sunny Sanyal","cross_cats":["cs.AI","stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-05T00:49:59Z","title":"Upweighting Easy Samples in Fine-Tuning Mitigates Forgetting"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.02797","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":"sha256:0d0fbf2ea4b055a0d2608ba82a7216d806bde13ba1229c5f1a97f9f352e1e49d","target":"record","created_at":"2026-07-05T11:20:08Z","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":"d4d85db667ced0e2c54af06eb0d9628ce27d76d265799e63cddbb7a2b96a67f4","cross_cats_sorted":["cs.AI","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-05T00:49:59Z","title_canon_sha256":"f2024857868a8db2102729c8a6a2a2645776bbc9f984e8cde1aa30be4a98965c"},"schema_version":"1.0","source":{"id":"2502.02797","kind":"arxiv","version":2}},"canonical_sha256":"1c9dbfaa8d501c87010daf944affa807311ce5018cae046a1c39c1166b6742f8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1c9dbfaa8d501c87010daf944affa807311ce5018cae046a1c39c1166b6742f8","first_computed_at":"2026-07-05T11:20:08.730485Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:20:08.730485Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Sra590f9nXzO8YT7myZwXplo80mVr8MGHwBKCMliJSj5/tZtaC8XJhDDfDXeQW+ZBrUutVPtz30kTBDAJEEqAA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:20:08.730958Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.02797","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0d0fbf2ea4b055a0d2608ba82a7216d806bde13ba1229c5f1a97f9f352e1e49d","sha256:a863776cbc56e4b4cec30b7df821bc9903bf019fa184133d504be3fa9ca3cdc2"],"state_sha256":"2f1221c236fda3e30d4e3bfa9f9ba2e2cd031bcc43aecf4498bfc8aebdfa599d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PeLEXz+c3cC+MY3561dhsoCOCnR/FjnL08/1MIkDqqS7WSf13m1gQh4U+OzlhZuXk/ng55Im7q08TMdBzspqBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T03:35:27.342649Z","bundle_sha256":"64857a672d804d58391e0e2833c3dd639528a8fad99d35be700ee3e1788a7371"}}