{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:WXBKJPT42PTT74S2GHDE6YPEQR","short_pith_number":"pith:WXBKJPT4","canonical_record":{"source":{"id":"2410.17413","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-22T20:39:21Z","cross_cats_sorted":[],"title_canon_sha256":"f8e84b5711653aabb4965ed3f8cd4e0c019eae5457c0a5fa1dad0de7a73b110c","abstract_canon_sha256":"51806af3407230c091b7c1ba52c986628eaaf39d6d2ffe60e47de3a2975e4501"},"schema_version":"1.0"},"canonical_sha256":"b5c2a4be7cd3e73ff25a31c64f61e48460bc7796c3473af45d4278357c51e7a1","source":{"kind":"arxiv","id":"2410.17413","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.17413","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"arxiv_version","alias_value":"2410.17413v3","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.17413","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"pith_short_12","alias_value":"WXBKJPT42PTT","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"pith_short_16","alias_value":"WXBKJPT42PTT74S2","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"pith_short_8","alias_value":"WXBKJPT4","created_at":"2026-07-05T09:52:40Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:WXBKJPT42PTT74S2GHDE6YPEQR","target":"record","payload":{"canonical_record":{"source":{"id":"2410.17413","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-22T20:39:21Z","cross_cats_sorted":[],"title_canon_sha256":"f8e84b5711653aabb4965ed3f8cd4e0c019eae5457c0a5fa1dad0de7a73b110c","abstract_canon_sha256":"51806af3407230c091b7c1ba52c986628eaaf39d6d2ffe60e47de3a2975e4501"},"schema_version":"1.0"},"canonical_sha256":"b5c2a4be7cd3e73ff25a31c64f61e48460bc7796c3473af45d4278357c51e7a1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:52:40.202108Z","signature_b64":"v9HgkBHlk0i0FOoVdp701i2bFLb+37IzH2zZJqMpjod0VjQSsQVJ7WJOYL8ywG6V2jAnbC10VDulqGO/ScQjAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b5c2a4be7cd3e73ff25a31c64f61e48460bc7796c3473af45d4278357c51e7a1","last_reissued_at":"2026-07-05T09:52:40.201686Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:52:40.201686Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.17413","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":"2026-07-05T09:52:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QYvsgbPGnga9WCgloNwH1KA+ySSKQE5aqBgNfUt797oIiuONm9oBhlIk1w+m3oQtkK7HgA4ErxGQRO/kH+c6Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T08:57:48.446572Z"},"content_sha256":"01b9077c49b79f5d59886bca223a5e853fd08ebe69b20792caf223557c499edb","schema_version":"1.0","event_id":"sha256:01b9077c49b79f5d59886bca223a5e853fd08ebe69b20792caf223557c499edb"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:WXBKJPT42PTT74S2GHDE6YPEQR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Scalable Influence and Fact Tracing for Large Language Model Pretraining","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dheeraj Rajagopal, Ian Tenney, Lucas Dixon, Tolga Bolukbasi, Tyler A. Chang","submitted_at":"2024-10-22T20:39:21Z","abstract_excerpt":"Training data attribution (TDA) methods aim to attribute model outputs back to specific training examples, and the application of these methods to large language model (LLM) outputs could significantly advance model transparency and data curation. However, it has been challenging to date to apply these methods to the full scale of LLM pretraining. In this paper, we refine existing gradient-based methods to work effectively at scale, allowing us to retrieve influential examples for an 8B-parameter language model from a pretraining corpus of over 160B tokens with no need for subsampling or pre-f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.17413","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/2410.17413/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":"2026-07-05T09:52:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MbEENV4DaK963SruXvTlzTLl1Kpeh+pWtnRgfZOgRwcbpSb+fubfrK48mNi7SKV8P1pQnrw8xLtI/TmJTqSzBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T08:57:48.446952Z"},"content_sha256":"b033507cae1727a7fd9c45baf22086a859b0a33f6147ce882f410b6fe954508c","schema_version":"1.0","event_id":"sha256:b033507cae1727a7fd9c45baf22086a859b0a33f6147ce882f410b6fe954508c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/WXBKJPT42PTT74S2GHDE6YPEQR/bundle.json","state_url":"https://pith.science/pith/WXBKJPT42PTT74S2GHDE6YPEQR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/WXBKJPT42PTT74S2GHDE6YPEQR/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":"2026-08-22T08:57:48Z","links":{"resolver":"https://pith.science/pith/WXBKJPT42PTT74S2GHDE6YPEQR","bundle":"https://pith.science/pith/WXBKJPT42PTT74S2GHDE6YPEQR/bundle.json","state":"https://pith.science/pith/WXBKJPT42PTT74S2GHDE6YPEQR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/WXBKJPT42PTT74S2GHDE6YPEQR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:WXBKJPT42PTT74S2GHDE6YPEQR","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":"51806af3407230c091b7c1ba52c986628eaaf39d6d2ffe60e47de3a2975e4501","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-22T20:39:21Z","title_canon_sha256":"f8e84b5711653aabb4965ed3f8cd4e0c019eae5457c0a5fa1dad0de7a73b110c"},"schema_version":"1.0","source":{"id":"2410.17413","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.17413","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"arxiv_version","alias_value":"2410.17413v3","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.17413","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"pith_short_12","alias_value":"WXBKJPT42PTT","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"pith_short_16","alias_value":"WXBKJPT42PTT74S2","created_at":"2026-07-05T09:52:40Z"},{"alias_kind":"pith_short_8","alias_value":"WXBKJPT4","created_at":"2026-07-05T09:52:40Z"}],"graph_snapshots":[{"event_id":"sha256:b033507cae1727a7fd9c45baf22086a859b0a33f6147ce882f410b6fe954508c","target":"graph","created_at":"2026-07-05T09:52:40Z","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/2410.17413/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training data attribution (TDA) methods aim to attribute model outputs back to specific training examples, and the application of these methods to large language model (LLM) outputs could significantly advance model transparency and data curation. However, it has been challenging to date to apply these methods to the full scale of LLM pretraining. In this paper, we refine existing gradient-based methods to work effectively at scale, allowing us to retrieve influential examples for an 8B-parameter language model from a pretraining corpus of over 160B tokens with no need for subsampling or pre-f","authors_text":"Dheeraj Rajagopal, Ian Tenney, Lucas Dixon, Tolga Bolukbasi, Tyler A. Chang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-22T20:39:21Z","title":"Scalable Influence and Fact Tracing for Large Language Model Pretraining"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.17413","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":"sha256:01b9077c49b79f5d59886bca223a5e853fd08ebe69b20792caf223557c499edb","target":"record","created_at":"2026-07-05T09:52:40Z","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":"51806af3407230c091b7c1ba52c986628eaaf39d6d2ffe60e47de3a2975e4501","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-22T20:39:21Z","title_canon_sha256":"f8e84b5711653aabb4965ed3f8cd4e0c019eae5457c0a5fa1dad0de7a73b110c"},"schema_version":"1.0","source":{"id":"2410.17413","kind":"arxiv","version":3}},"canonical_sha256":"b5c2a4be7cd3e73ff25a31c64f61e48460bc7796c3473af45d4278357c51e7a1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b5c2a4be7cd3e73ff25a31c64f61e48460bc7796c3473af45d4278357c51e7a1","first_computed_at":"2026-07-05T09:52:40.201686Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:52:40.201686Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"v9HgkBHlk0i0FOoVdp701i2bFLb+37IzH2zZJqMpjod0VjQSsQVJ7WJOYL8ywG6V2jAnbC10VDulqGO/ScQjAg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:52:40.202108Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.17413","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:01b9077c49b79f5d59886bca223a5e853fd08ebe69b20792caf223557c499edb","sha256:b033507cae1727a7fd9c45baf22086a859b0a33f6147ce882f410b6fe954508c"],"state_sha256":"de97f16060da727e95b59a2a1bb6fbd800689cf787bb9aefde84bf16603003c8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"txLAiQ/kEuW/mpE+umm0Xu0xlR81+mAUFXDhNnijoWangEQqSGkcDwqoFFW4hb5Y0OBft2VqK9V58KuAhKEzCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-22T08:57:48.450123Z","bundle_sha256":"136301a60b16ffb7daa427a994446d7d748109b808b2d3a311925f3339e1f70c"}}