{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:6US5UUSVFC46X6CMTD6OBMLEBY","short_pith_number":"pith:6US5UUSV","canonical_record":{"source":{"id":"2508.01908","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-03T20:07:15Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"91ba8f78e45dbe0f77517341bdbd3027e23918972cd1aebcf3fe5fce52348124","abstract_canon_sha256":"c6beca9fc1fde5389783e7f4bc6381dfa322434769cea4d688bdbd5a4582a722"},"schema_version":"1.0"},"canonical_sha256":"f525da525528b9ebf84c98fce0b1640e3e5746f36deeb1f31a8a228bf6351d29","source":{"kind":"arxiv","id":"2508.01908","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.01908","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"arxiv_version","alias_value":"2508.01908v1","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.01908","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"pith_short_12","alias_value":"6US5UUSVFC46","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"pith_short_16","alias_value":"6US5UUSVFC46X6CM","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"pith_short_8","alias_value":"6US5UUSV","created_at":"2026-07-05T11:47:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:6US5UUSVFC46X6CMTD6OBMLEBY","target":"record","payload":{"canonical_record":{"source":{"id":"2508.01908","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-03T20:07:15Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"91ba8f78e45dbe0f77517341bdbd3027e23918972cd1aebcf3fe5fce52348124","abstract_canon_sha256":"c6beca9fc1fde5389783e7f4bc6381dfa322434769cea4d688bdbd5a4582a722"},"schema_version":"1.0"},"canonical_sha256":"f525da525528b9ebf84c98fce0b1640e3e5746f36deeb1f31a8a228bf6351d29","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:47:47.988859Z","signature_b64":"SE8VWgiYi0lkWnwQvnJy10GCfV+GYXHdH2bZguzGklCfCM9NdYguC93TXBOZvYPqu75GY0XxxLBvJ6YcTdEsBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f525da525528b9ebf84c98fce0b1640e3e5746f36deeb1f31a8a228bf6351d29","last_reissued_at":"2026-07-05T11:47:47.988379Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:47:47.988379Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.01908","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:47:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dW9Vnv22adfyuShPpGVa+ZmePveFG38n9aFLO9O/jAd1sViyw3yy39TlCHdckxQ5pPcE1PQGPI7CKnnRTixvDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-13T01:01:46.112474Z"},"content_sha256":"76e78f26c80b02cacbd203a769c6e255346d732e8c57f02238a35bf7c176b732","schema_version":"1.0","event_id":"sha256:76e78f26c80b02cacbd203a769c6e255346d732e8c57f02238a35bf7c176b732"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:6US5UUSVFC46X6CMTD6OBMLEBY","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Revisiting Replay and Gradient Alignment for Continual Pre-Training of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Benjamin Therien, Gopeshh Subbaraj, Hiroaki Kingetsu, Irina Rish, Istabrak Abbes, Matthew Riemer, Nizar Islah, Sarath Chandar, Tsuguchika Tabaru","submitted_at":"2025-08-03T20:07:15Z","abstract_excerpt":"Training large language models (LLMs) typically involves pre-training on massive corpora, only to restart the process entirely when new data becomes available. A more efficient and resource-conserving approach would be continual pre-training, where models are updated with new data rather than retraining from scratch. However, the introduction of new data often causes distribution shifts, leading to performance degradation on previously learned tasks. In this paper, we take a deeper look at two popular proposals for addressing this distribution shift within the continual learning literature: ex"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.01908","kind":"arxiv","version":1},"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/2508.01908/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":"2026-07-05T11:47:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"yqQq+u5skO3VbCUz+KYjOBh7FVtTHzGDIc/5gSkdGugYdJyyNDTtf1ZHffqJ5cSTf2n4BAJ24VHIbf7CamNcDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-13T01:01:46.113103Z"},"content_sha256":"eb2f8e9e64314b25ddaf4fd0cef985caf6dba800fbd047035b64277c79e28c09","schema_version":"1.0","event_id":"sha256:eb2f8e9e64314b25ddaf4fd0cef985caf6dba800fbd047035b64277c79e28c09"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6US5UUSVFC46X6CMTD6OBMLEBY/bundle.json","state_url":"https://pith.science/pith/6US5UUSVFC46X6CMTD6OBMLEBY/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6US5UUSVFC46X6CMTD6OBMLEBY/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":"2026-07-13T01:01:46Z","links":{"resolver":"https://pith.science/pith/6US5UUSVFC46X6CMTD6OBMLEBY","bundle":"https://pith.science/pith/6US5UUSVFC46X6CMTD6OBMLEBY/bundle.json","state":"https://pith.science/pith/6US5UUSVFC46X6CMTD6OBMLEBY/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6US5UUSVFC46X6CMTD6OBMLEBY/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:6US5UUSVFC46X6CMTD6OBMLEBY","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":"c6beca9fc1fde5389783e7f4bc6381dfa322434769cea4d688bdbd5a4582a722","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-03T20:07:15Z","title_canon_sha256":"91ba8f78e45dbe0f77517341bdbd3027e23918972cd1aebcf3fe5fce52348124"},"schema_version":"1.0","source":{"id":"2508.01908","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.01908","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"arxiv_version","alias_value":"2508.01908v1","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.01908","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"pith_short_12","alias_value":"6US5UUSVFC46","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"pith_short_16","alias_value":"6US5UUSVFC46X6CM","created_at":"2026-07-05T11:47:47Z"},{"alias_kind":"pith_short_8","alias_value":"6US5UUSV","created_at":"2026-07-05T11:47:47Z"}],"graph_snapshots":[{"event_id":"sha256:eb2f8e9e64314b25ddaf4fd0cef985caf6dba800fbd047035b64277c79e28c09","target":"graph","created_at":"2026-07-05T11:47:47Z","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/2508.01908/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training large language models (LLMs) typically involves pre-training on massive corpora, only to restart the process entirely when new data becomes available. A more efficient and resource-conserving approach would be continual pre-training, where models are updated with new data rather than retraining from scratch. However, the introduction of new data often causes distribution shifts, leading to performance degradation on previously learned tasks. In this paper, we take a deeper look at two popular proposals for addressing this distribution shift within the continual learning literature: ex","authors_text":"Benjamin Therien, Gopeshh Subbaraj, Hiroaki Kingetsu, Irina Rish, Istabrak Abbes, Matthew Riemer, Nizar Islah, Sarath Chandar, Tsuguchika Tabaru","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-03T20:07:15Z","title":"Revisiting Replay and Gradient Alignment for Continual Pre-Training of Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.01908","kind":"arxiv","version":1},"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":"sha256:76e78f26c80b02cacbd203a769c6e255346d732e8c57f02238a35bf7c176b732","target":"record","created_at":"2026-07-05T11:47:47Z","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":"c6beca9fc1fde5389783e7f4bc6381dfa322434769cea4d688bdbd5a4582a722","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-03T20:07:15Z","title_canon_sha256":"91ba8f78e45dbe0f77517341bdbd3027e23918972cd1aebcf3fe5fce52348124"},"schema_version":"1.0","source":{"id":"2508.01908","kind":"arxiv","version":1}},"canonical_sha256":"f525da525528b9ebf84c98fce0b1640e3e5746f36deeb1f31a8a228bf6351d29","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f525da525528b9ebf84c98fce0b1640e3e5746f36deeb1f31a8a228bf6351d29","first_computed_at":"2026-07-05T11:47:47.988379Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:47:47.988379Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SE8VWgiYi0lkWnwQvnJy10GCfV+GYXHdH2bZguzGklCfCM9NdYguC93TXBOZvYPqu75GY0XxxLBvJ6YcTdEsBg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:47:47.988859Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.01908","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:76e78f26c80b02cacbd203a769c6e255346d732e8c57f02238a35bf7c176b732","sha256:eb2f8e9e64314b25ddaf4fd0cef985caf6dba800fbd047035b64277c79e28c09"],"state_sha256":"9197675f400ee85b4e1a4c53d1d59010fecee6fa0b8cebb6d0f97aa7bf19c4cb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WJ8yp383tSreej1Gst/gZAqfoI7PxPxPp8mmytcX2GHDkrShlpb2OEj02mo13YjjgiVIWwLWI7ppSYbvxeC4Bw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-13T01:01:46.115829Z","bundle_sha256":"9b3f338a01f0685e8f9ce7a0f9016536acd6fda0f62e1d5db985d196e3684555"}}