{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:ZKR32REINXSWTTGNIT5IJCY6KR","short_pith_number":"pith:ZKR32REI","canonical_record":{"source":{"id":"2505.07796","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-12T17:47:32Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"fa34be51d8524371435b338200fa645ca3bbdff8144668a89369c343e5c915bb","abstract_canon_sha256":"84760f40d4aa89125b55a04c1656ad6b2673c0b4b7985de80df2c20ed7f52e80"},"schema_version":"1.0"},"canonical_sha256":"caa3bd44886de569cccd44fa848b1e547d881542a81400d13c98c6b306f18d3d","source":{"kind":"arxiv","id":"2505.07796","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.07796","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"arxiv_version","alias_value":"2505.07796v2","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.07796","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"pith_short_12","alias_value":"ZKR32REINXSW","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"pith_short_16","alias_value":"ZKR32REINXSWTTGN","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"pith_short_8","alias_value":"ZKR32REI","created_at":"2026-07-05T11:24:20Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:ZKR32REINXSWTTGNIT5IJCY6KR","target":"record","payload":{"canonical_record":{"source":{"id":"2505.07796","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-12T17:47:32Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"fa34be51d8524371435b338200fa645ca3bbdff8144668a89369c343e5c915bb","abstract_canon_sha256":"84760f40d4aa89125b55a04c1656ad6b2673c0b4b7985de80df2c20ed7f52e80"},"schema_version":"1.0"},"canonical_sha256":"caa3bd44886de569cccd44fa848b1e547d881542a81400d13c98c6b306f18d3d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:24:20.063508Z","signature_b64":"mUWzrv0Z1TUh88KhW+tRF24DzAZw6QqMMe8gJMxct+A63Pez0ScHknCzgaIPU2aDWxCdSKvFRay6HyEauwCdAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"caa3bd44886de569cccd44fa848b1e547d881542a81400d13c98c6b306f18d3d","last_reissued_at":"2026-07-05T11:24:20.062981Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:24:20.062981Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.07796","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":"2026-07-05T11:24:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"m7CQe8+Ha5yg9337QyIPZWHwcumnE0Q+RNBCpRyiNAlZoOeWKAOW4zamhkVTyrC2t6SWjW+UTuRSSs/+Ug0fAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T05:14:19.417525Z"},"content_sha256":"4c857c8fe81db051bd79b7ea15674e433d7844f1c2c349ea9d7f3d68553ede77","schema_version":"1.0","event_id":"sha256:4c857c8fe81db051bd79b7ea15674e433d7844f1c2c349ea9d7f3d68553ede77"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:ZKR32REINXSWTTGNIT5IJCY6KR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning Dynamics in Continual Pre-Training for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Daniel Dajun Zeng, Howe Tissue, Linjing Li, Lu Wang, Xingjin Wang","submitted_at":"2025-05-12T17:47:32Z","abstract_excerpt":"Continual Pre-Training (CPT) has become a popular and effective method to apply strong foundation models to specific downstream tasks. In this work, we explore the learning dynamics throughout the CPT process for large language models. We specifically focus on how general and downstream domain performance evolves at each training step, with domain performance measured via validation losses. We have observed that the CPT loss curve fundamentally characterizes the transition from one curve to another hidden curve, and could be described by decoupling the effects of distribution shift and learnin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.07796","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/2505.07796/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":"2026-07-05T11:24:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"OOv8fchIWNwhPmadvAxr8HyjmYld+3E/k/eH4r3PI01mkwYZUTu23FZgu7YrREGSwbiTEygGxhcyPbLorjVZBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T05:14:19.418030Z"},"content_sha256":"6e481f0b81db0cc7cf34895a879e250bcc10e164e3cc55b41f5fbbecd68fec3a","schema_version":"1.0","event_id":"sha256:6e481f0b81db0cc7cf34895a879e250bcc10e164e3cc55b41f5fbbecd68fec3a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZKR32REINXSWTTGNIT5IJCY6KR/bundle.json","state_url":"https://pith.science/pith/ZKR32REINXSWTTGNIT5IJCY6KR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZKR32REINXSWTTGNIT5IJCY6KR/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":"2026-08-15T05:14:19Z","links":{"resolver":"https://pith.science/pith/ZKR32REINXSWTTGNIT5IJCY6KR","bundle":"https://pith.science/pith/ZKR32REINXSWTTGNIT5IJCY6KR/bundle.json","state":"https://pith.science/pith/ZKR32REINXSWTTGNIT5IJCY6KR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZKR32REINXSWTTGNIT5IJCY6KR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:ZKR32REINXSWTTGNIT5IJCY6KR","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":"84760f40d4aa89125b55a04c1656ad6b2673c0b4b7985de80df2c20ed7f52e80","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-12T17:47:32Z","title_canon_sha256":"fa34be51d8524371435b338200fa645ca3bbdff8144668a89369c343e5c915bb"},"schema_version":"1.0","source":{"id":"2505.07796","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.07796","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"arxiv_version","alias_value":"2505.07796v2","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.07796","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"pith_short_12","alias_value":"ZKR32REINXSW","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"pith_short_16","alias_value":"ZKR32REINXSWTTGN","created_at":"2026-07-05T11:24:20Z"},{"alias_kind":"pith_short_8","alias_value":"ZKR32REI","created_at":"2026-07-05T11:24:20Z"}],"graph_snapshots":[{"event_id":"sha256:6e481f0b81db0cc7cf34895a879e250bcc10e164e3cc55b41f5fbbecd68fec3a","target":"graph","created_at":"2026-07-05T11:24:20Z","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/2505.07796/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Continual Pre-Training (CPT) has become a popular and effective method to apply strong foundation models to specific downstream tasks. In this work, we explore the learning dynamics throughout the CPT process for large language models. We specifically focus on how general and downstream domain performance evolves at each training step, with domain performance measured via validation losses. We have observed that the CPT loss curve fundamentally characterizes the transition from one curve to another hidden curve, and could be described by decoupling the effects of distribution shift and learnin","authors_text":"Daniel Dajun Zeng, Howe Tissue, Linjing Li, Lu Wang, Xingjin Wang","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-12T17:47:32Z","title":"Learning Dynamics in Continual Pre-Training for Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.07796","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":"sha256:4c857c8fe81db051bd79b7ea15674e433d7844f1c2c349ea9d7f3d68553ede77","target":"record","created_at":"2026-07-05T11:24:20Z","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":"84760f40d4aa89125b55a04c1656ad6b2673c0b4b7985de80df2c20ed7f52e80","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-12T17:47:32Z","title_canon_sha256":"fa34be51d8524371435b338200fa645ca3bbdff8144668a89369c343e5c915bb"},"schema_version":"1.0","source":{"id":"2505.07796","kind":"arxiv","version":2}},"canonical_sha256":"caa3bd44886de569cccd44fa848b1e547d881542a81400d13c98c6b306f18d3d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"caa3bd44886de569cccd44fa848b1e547d881542a81400d13c98c6b306f18d3d","first_computed_at":"2026-07-05T11:24:20.062981Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:24:20.062981Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"mUWzrv0Z1TUh88KhW+tRF24DzAZw6QqMMe8gJMxct+A63Pez0ScHknCzgaIPU2aDWxCdSKvFRay6HyEauwCdAw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:24:20.063508Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.07796","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4c857c8fe81db051bd79b7ea15674e433d7844f1c2c349ea9d7f3d68553ede77","sha256:6e481f0b81db0cc7cf34895a879e250bcc10e164e3cc55b41f5fbbecd68fec3a"],"state_sha256":"b773014d9e1c6d6d6f1a069e3335f8a2a2876a7acdfcfd302a975e8a961a756f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tUAM7tiCMC+Zu3CRFOQWJtxGgqYa2eKnJtGB3GiqJkVokiAnmfSGeXGrIgUfB/S1Ud+h5hXsCUmLqf3KNM3xBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-15T05:14:19.423080Z","bundle_sha256":"a5c551c9482889c448030b44f1d3c7666ee8fc5bfa12b94b75647cfedfbcb962"}}