{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:UM7G25HQOPT4M2ODM64MYTXT3S","short_pith_number":"pith:UM7G25HQ","canonical_record":{"source":{"id":"2402.02868","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T10:30:47Z","cross_cats_sorted":[],"title_canon_sha256":"cae6673199a9b474e819114cebb12f947504aa0fabe05ff86efb20637d2519e0","abstract_canon_sha256":"7633a0f434f2d07949ab1d2fad817e954142b55090928070f902c536c5e48109"},"schema_version":"1.0"},"canonical_sha256":"a33e6d74f073e7c669c367b8cc4ef3dcb7cdf163d2a5d25c619343d346e42ea0","source":{"kind":"arxiv","id":"2402.02868","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.02868","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"arxiv_version","alias_value":"2402.02868v3","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.02868","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"pith_short_12","alias_value":"UM7G25HQOPT4","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"pith_short_16","alias_value":"UM7G25HQOPT4M2OD","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"pith_short_8","alias_value":"UM7G25HQ","created_at":"2026-07-05T08:44:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:UM7G25HQOPT4M2ODM64MYTXT3S","target":"record","payload":{"canonical_record":{"source":{"id":"2402.02868","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T10:30:47Z","cross_cats_sorted":[],"title_canon_sha256":"cae6673199a9b474e819114cebb12f947504aa0fabe05ff86efb20637d2519e0","abstract_canon_sha256":"7633a0f434f2d07949ab1d2fad817e954142b55090928070f902c536c5e48109"},"schema_version":"1.0"},"canonical_sha256":"a33e6d74f073e7c669c367b8cc4ef3dcb7cdf163d2a5d25c619343d346e42ea0","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:44:50.810781Z","signature_b64":"5IJm1r3gYsE/p5lrKxeZKr7NmWd7hbJ0sT5V2hciUc/VcRmXTDax+XLUvYGPmTQeBe/oc8LiXAgRyRPnkoPEBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a33e6d74f073e7c669c367b8cc4ef3dcb7cdf163d2a5d25c619343d346e42ea0","last_reissued_at":"2026-07-05T08:44:50.810351Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:44:50.810351Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.02868","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":"2026-07-05T08:44:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VMFwRY3hj7EWAPie0Y885WHRe1e/TilYGeg9O9HCBfVfuAhEiyk3Y/CJrvlWtaYbsk6pYGoVvPonoT554byaDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-20T10:53:41.875932Z"},"content_sha256":"7c44d8f5f7c782748196750b560bd3610fa3dab084ac49d7b0e16eec6fc08d8e","schema_version":"1.0","event_id":"sha256:7c44d8f5f7c782748196750b560bd3610fa3dab084ac49d7b0e16eec6fc08d8e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:UM7G25HQOPT4M2ODM64MYTXT3S","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Fine-tuning Reinforcement Learning Models is Secretly a Forgetting Mitigation Problem","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bart{\\l}omiej Cupia{\\l}, {\\L}ukasz Kuci\\'nski, Maciej Wo{\\l}czyk, Mateusz Ostaszewski, Micha{\\l} Bortkiewicz, Micha{\\l} Zaj\\k{a}c, Piotr Mi{\\l}o\\'s, Razvan Pascanu","submitted_at":"2024-02-05T10:30:47Z","abstract_excerpt":"Fine-tuning is a widespread technique that allows practitioners to transfer pre-trained capabilities, as recently showcased by the successful applications of foundation models. However, fine-tuning reinforcement learning (RL) models remains a challenge. This work conceptualizes one specific cause of poor transfer, accentuated in the RL setting by the interplay between actions and observations: forgetting of pre-trained capabilities. Namely, a model deteriorates on the state subspace of the downstream task not visited in the initial phase of fine-tuning, on which the model behaved well due to p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.02868","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/2402.02868/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":"2026-07-05T08:44:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MKMafZRPbgckHHGckcqsgs58jMJ3HOYjfOBm3uZhHTgN8IuiJzQ7otQTMOYYFJV6f7P41YEMN9hLWIAd5fQEBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-20T10:53:41.876461Z"},"content_sha256":"ca5478cb7271f6c1188c5e4a27cdb5ce7839c8e555786fdbfb8d2e8499c415c0","schema_version":"1.0","event_id":"sha256:ca5478cb7271f6c1188c5e4a27cdb5ce7839c8e555786fdbfb8d2e8499c415c0"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/UM7G25HQOPT4M2ODM64MYTXT3S/bundle.json","state_url":"https://pith.science/pith/UM7G25HQOPT4M2ODM64MYTXT3S/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/UM7G25HQOPT4M2ODM64MYTXT3S/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":"2026-08-20T10:53:41Z","links":{"resolver":"https://pith.science/pith/UM7G25HQOPT4M2ODM64MYTXT3S","bundle":"https://pith.science/pith/UM7G25HQOPT4M2ODM64MYTXT3S/bundle.json","state":"https://pith.science/pith/UM7G25HQOPT4M2ODM64MYTXT3S/state.json","well_known_bundle":"https://pith.science/.well-known/pith/UM7G25HQOPT4M2ODM64MYTXT3S/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:UM7G25HQOPT4M2ODM64MYTXT3S","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":"7633a0f434f2d07949ab1d2fad817e954142b55090928070f902c536c5e48109","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T10:30:47Z","title_canon_sha256":"cae6673199a9b474e819114cebb12f947504aa0fabe05ff86efb20637d2519e0"},"schema_version":"1.0","source":{"id":"2402.02868","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.02868","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"arxiv_version","alias_value":"2402.02868v3","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.02868","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"pith_short_12","alias_value":"UM7G25HQOPT4","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"pith_short_16","alias_value":"UM7G25HQOPT4M2OD","created_at":"2026-07-05T08:44:50Z"},{"alias_kind":"pith_short_8","alias_value":"UM7G25HQ","created_at":"2026-07-05T08:44:50Z"}],"graph_snapshots":[{"event_id":"sha256:ca5478cb7271f6c1188c5e4a27cdb5ce7839c8e555786fdbfb8d2e8499c415c0","target":"graph","created_at":"2026-07-05T08:44:50Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2402.02868/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Fine-tuning is a widespread technique that allows practitioners to transfer pre-trained capabilities, as recently showcased by the successful applications of foundation models. However, fine-tuning reinforcement learning (RL) models remains a challenge. This work conceptualizes one specific cause of poor transfer, accentuated in the RL setting by the interplay between actions and observations: forgetting of pre-trained capabilities. Namely, a model deteriorates on the state subspace of the downstream task not visited in the initial phase of fine-tuning, on which the model behaved well due to p","authors_text":"Bart{\\l}omiej Cupia{\\l}, {\\L}ukasz Kuci\\'nski, Maciej Wo{\\l}czyk, Mateusz Ostaszewski, Micha{\\l} Bortkiewicz, Micha{\\l} Zaj\\k{a}c, Piotr Mi{\\l}o\\'s, Razvan Pascanu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T10:30:47Z","title":"Fine-tuning Reinforcement Learning Models is Secretly a Forgetting Mitigation Problem"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.02868","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":"sha256:7c44d8f5f7c782748196750b560bd3610fa3dab084ac49d7b0e16eec6fc08d8e","target":"record","created_at":"2026-07-05T08:44:50Z","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":"7633a0f434f2d07949ab1d2fad817e954142b55090928070f902c536c5e48109","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T10:30:47Z","title_canon_sha256":"cae6673199a9b474e819114cebb12f947504aa0fabe05ff86efb20637d2519e0"},"schema_version":"1.0","source":{"id":"2402.02868","kind":"arxiv","version":3}},"canonical_sha256":"a33e6d74f073e7c669c367b8cc4ef3dcb7cdf163d2a5d25c619343d346e42ea0","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a33e6d74f073e7c669c367b8cc4ef3dcb7cdf163d2a5d25c619343d346e42ea0","first_computed_at":"2026-07-05T08:44:50.810351Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:44:50.810351Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"5IJm1r3gYsE/p5lrKxeZKr7NmWd7hbJ0sT5V2hciUc/VcRmXTDax+XLUvYGPmTQeBe/oc8LiXAgRyRPnkoPEBA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:44:50.810781Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.02868","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7c44d8f5f7c782748196750b560bd3610fa3dab084ac49d7b0e16eec6fc08d8e","sha256:ca5478cb7271f6c1188c5e4a27cdb5ce7839c8e555786fdbfb8d2e8499c415c0"],"state_sha256":"e6e770937f32a0b28208012992248ef239d067cd10c3005ec9d83a78b76b400a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lpExQrOwpTGHXjOklmLuCWwmK3/jDDNszwskZGjBHkMNAcVIx52ZYjFXN9/iKMHApGuqG/E0xLlg+doSajpIAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-20T10:53:41.881785Z","bundle_sha256":"609cc46d14b68b0ffe3b3b23bb660a5a7d3e7dd055df48bbfc3a0f8a8ee96b51"}}