{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:SP22TY2PMTFJMW42EBIYT3XPPT","short_pith_number":"pith:SP22TY2P","canonical_record":{"source":{"id":"2409.05816","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-09T17:23:29Z","cross_cats_sorted":["cs.LG","stat.ML"],"title_canon_sha256":"a675701def5eaaab2ae50b2e3bb3cde31d3bfb23e5902aa8947758c51630e321","abstract_canon_sha256":"286cda8fdc35c9d38dcc2e37f7736b18a12a27f1379897f5e02b953a9dc5cf22"},"schema_version":"1.0"},"canonical_sha256":"93f5a9e34f64ca965b9a205189eeef7cc5f170135845c3700170f5e593295d81","source":{"kind":"arxiv","id":"2409.05816","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2409.05816","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"arxiv_version","alias_value":"2409.05816v2","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.05816","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_12","alias_value":"SP22TY2PMTFJ","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_16","alias_value":"SP22TY2PMTFJMW42","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_8","alias_value":"SP22TY2P","created_at":"2026-07-05T10:28:05Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:SP22TY2PMTFJMW42EBIYT3XPPT","target":"record","payload":{"canonical_record":{"source":{"id":"2409.05816","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-09T17:23:29Z","cross_cats_sorted":["cs.LG","stat.ML"],"title_canon_sha256":"a675701def5eaaab2ae50b2e3bb3cde31d3bfb23e5902aa8947758c51630e321","abstract_canon_sha256":"286cda8fdc35c9d38dcc2e37f7736b18a12a27f1379897f5e02b953a9dc5cf22"},"schema_version":"1.0"},"canonical_sha256":"93f5a9e34f64ca965b9a205189eeef7cc5f170135845c3700170f5e593295d81","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:28:05.018858Z","signature_b64":"hpHC9IwnoaQpJmuycVGhhNh4OUeTSq29ZxMD7n+13dsBnGvdKKdg2FnWn46aEZbHyfjF1MavMfbYlacDlCG1Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"93f5a9e34f64ca965b9a205189eeef7cc5f170135845c3700170f5e593295d81","last_reissued_at":"2026-07-05T10:28:05.017891Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:28:05.017891Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2409.05816","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":"2026-07-05T10:28:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/luz1rybHeX0Uq6HeMAsEuuQ5rIx+aGA4/b3fszGcWrUs//6KONm8h8ADDdyZHXw+DcngQ3Qe/t/Ff0Nu+SqBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T18:36:31.586458Z"},"content_sha256":"328aa07408de30a51d4d738a85b81fe24ec857bee68fdc48ce4edf79c5b62274","schema_version":"1.0","event_id":"sha256:328aa07408de30a51d4d738a85b81fe24ec857bee68fdc48ce4edf79c5b62274"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:SP22TY2PMTFJMW42EBIYT3XPPT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Improving Pretraining Data Using Perplexity Correlations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Christopher Potts, Tatsunori Hashimoto, Tristan Thrush","submitted_at":"2024-09-09T17:23:29Z","abstract_excerpt":"Quality pretraining data is often seen as the key to high-performance language models. However, progress in understanding pretraining data has been slow due to the costly pretraining runs required for data selection experiments. We present a framework that avoids these costs and selects high-quality pretraining data without any LLM training of our own. Our work is based on a simple observation: LLM losses on many pretraining texts are correlated with downstream benchmark performance, and selecting high-correlation documents is an effective pretraining data selection method. We build a new stat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.05816","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/2409.05816/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":"2026-07-05T10:28:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZNFLIopErQkxMINhhyENOe76sG9z9pP1MaNvCc4flcubClNxkMDMbPvEysiFOJe3YWLDu0w4b82MPJyufq1tDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T18:36:31.587303Z"},"content_sha256":"36c282761dae9830643cc394641db8b43c44fff0a1a31a6fdf551a1eb686abe1","schema_version":"1.0","event_id":"sha256:36c282761dae9830643cc394641db8b43c44fff0a1a31a6fdf551a1eb686abe1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SP22TY2PMTFJMW42EBIYT3XPPT/bundle.json","state_url":"https://pith.science/pith/SP22TY2PMTFJMW42EBIYT3XPPT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SP22TY2PMTFJMW42EBIYT3XPPT/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":"2026-08-14T18:36:31Z","links":{"resolver":"https://pith.science/pith/SP22TY2PMTFJMW42EBIYT3XPPT","bundle":"https://pith.science/pith/SP22TY2PMTFJMW42EBIYT3XPPT/bundle.json","state":"https://pith.science/pith/SP22TY2PMTFJMW42EBIYT3XPPT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SP22TY2PMTFJMW42EBIYT3XPPT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:SP22TY2PMTFJMW42EBIYT3XPPT","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":"286cda8fdc35c9d38dcc2e37f7736b18a12a27f1379897f5e02b953a9dc5cf22","cross_cats_sorted":["cs.LG","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-09T17:23:29Z","title_canon_sha256":"a675701def5eaaab2ae50b2e3bb3cde31d3bfb23e5902aa8947758c51630e321"},"schema_version":"1.0","source":{"id":"2409.05816","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2409.05816","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"arxiv_version","alias_value":"2409.05816v2","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.05816","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_12","alias_value":"SP22TY2PMTFJ","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_16","alias_value":"SP22TY2PMTFJMW42","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_8","alias_value":"SP22TY2P","created_at":"2026-07-05T10:28:05Z"}],"graph_snapshots":[{"event_id":"sha256:36c282761dae9830643cc394641db8b43c44fff0a1a31a6fdf551a1eb686abe1","target":"graph","created_at":"2026-07-05T10:28:05Z","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/2409.05816/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Quality pretraining data is often seen as the key to high-performance language models. However, progress in understanding pretraining data has been slow due to the costly pretraining runs required for data selection experiments. We present a framework that avoids these costs and selects high-quality pretraining data without any LLM training of our own. Our work is based on a simple observation: LLM losses on many pretraining texts are correlated with downstream benchmark performance, and selecting high-correlation documents is an effective pretraining data selection method. We build a new stat","authors_text":"Christopher Potts, Tatsunori Hashimoto, Tristan Thrush","cross_cats":["cs.LG","stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-09T17:23:29Z","title":"Improving Pretraining Data Using Perplexity Correlations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.05816","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":"sha256:328aa07408de30a51d4d738a85b81fe24ec857bee68fdc48ce4edf79c5b62274","target":"record","created_at":"2026-07-05T10:28:05Z","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":"286cda8fdc35c9d38dcc2e37f7736b18a12a27f1379897f5e02b953a9dc5cf22","cross_cats_sorted":["cs.LG","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-09T17:23:29Z","title_canon_sha256":"a675701def5eaaab2ae50b2e3bb3cde31d3bfb23e5902aa8947758c51630e321"},"schema_version":"1.0","source":{"id":"2409.05816","kind":"arxiv","version":2}},"canonical_sha256":"93f5a9e34f64ca965b9a205189eeef7cc5f170135845c3700170f5e593295d81","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"93f5a9e34f64ca965b9a205189eeef7cc5f170135845c3700170f5e593295d81","first_computed_at":"2026-07-05T10:28:05.017891Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:28:05.017891Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"hpHC9IwnoaQpJmuycVGhhNh4OUeTSq29ZxMD7n+13dsBnGvdKKdg2FnWn46aEZbHyfjF1MavMfbYlacDlCG1Ag==","signature_status":"signed_v1","signed_at":"2026-07-05T10:28:05.018858Z","signed_message":"canonical_sha256_bytes"},"source_id":"2409.05816","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:328aa07408de30a51d4d738a85b81fe24ec857bee68fdc48ce4edf79c5b62274","sha256:36c282761dae9830643cc394641db8b43c44fff0a1a31a6fdf551a1eb686abe1"],"state_sha256":"bd267d704e373c4006ec2db97c077f67251fff93ad6e0cc5f082a7b996480410"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9rUTtpGPcwW1FHU5Lw3BOc68dVaHO/oe2KL+8/YhB2bpEWl/jUYD9wa74dNoTUhD3pBgeTTg7ku6j+O54cMRAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-14T18:36:31.592711Z","bundle_sha256":"4d9f309c2091b48e3b385cfac57cad191b038cf3becc5b7533fc6620ec053309"}}