{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NLMEDV5DKMJQ7JA2VCX4HAR7GV","short_pith_number":"pith:NLMEDV5D","schema_version":"1.0","canonical_sha256":"6ad841d7a353130fa41aa8afc3823f356f19f365fba7ce126d22cafcae09f10e","source":{"kind":"arxiv","id":"2405.01525","version":1},"attestation_state":"computed","paper":{"title":"FLAME: Factuality-Aware Alignment for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Barlas Oguz, Jimmy Lin, Luyu Gao, Sheng-Chieh Lin, Wenhan Xiong, Wen-tau Yih, Xilun Chen","submitted_at":"2024-05-02T17:54:54Z","abstract_excerpt":"Alignment is a standard procedure to fine-tune pre-trained large language models (LLMs) to follow natural language instructions and serve as helpful AI assistants. We have observed, however, that the conventional alignment process fails to enhance the factual accuracy of LLMs, and often leads to the generation of more false facts (i.e. hallucination). In this paper, we study how to make the LLM alignment process more factual, by first identifying factors that lead to hallucination in both alignment steps:\\ supervised fine-tuning (SFT) and reinforcement learning (RL). In particular, we find tha"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2405.01525","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-02T17:54:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"731e998ced7ac89b764ae487f63f121b2d9c5a289969caa7ddaa57170b3fad14","abstract_canon_sha256":"f3bfaff671292c3955f8bd4b3d5f9417fa2a28025a6dd3f6325b7785b308b834"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:14:46.368462Z","signature_b64":"xEdJlOb0iARa4I3uLq5KG09lIpX6Mb9Pt4Bv3/ZgGe2FPEvkHPJjD+vnQPMwGpBPmhR+DjV1S7tI/8IefyWRAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ad841d7a353130fa41aa8afc3823f356f19f365fba7ce126d22cafcae09f10e","last_reissued_at":"2026-07-05T08:14:46.368007Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:14:46.368007Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FLAME: Factuality-Aware Alignment for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Barlas Oguz, Jimmy Lin, Luyu Gao, Sheng-Chieh Lin, Wenhan Xiong, Wen-tau Yih, Xilun Chen","submitted_at":"2024-05-02T17:54:54Z","abstract_excerpt":"Alignment is a standard procedure to fine-tune pre-trained large language models (LLMs) to follow natural language instructions and serve as helpful AI assistants. We have observed, however, that the conventional alignment process fails to enhance the factual accuracy of LLMs, and often leads to the generation of more false facts (i.e. hallucination). In this paper, we study how to make the LLM alignment process more factual, by first identifying factors that lead to hallucination in both alignment steps:\\ supervised fine-tuning (SFT) and reinforcement learning (RL). In particular, we find tha"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.01525","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/2405.01525/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2405.01525","created_at":"2026-07-05T08:14:46.368060+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.01525v1","created_at":"2026-07-05T08:14:46.368060+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.01525","created_at":"2026-07-05T08:14:46.368060+00:00"},{"alias_kind":"pith_short_12","alias_value":"NLMEDV5DKMJQ","created_at":"2026-07-05T08:14:46.368060+00:00"},{"alias_kind":"pith_short_16","alias_value":"NLMEDV5DKMJQ7JA2","created_at":"2026-07-05T08:14:46.368060+00:00"},{"alias_kind":"pith_short_8","alias_value":"NLMEDV5D","created_at":"2026-07-05T08:14:46.368060+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.05288","citing_title":"A Survey on Proactive Defense Strategies Against Misinformation in Large Language Models","ref_index":39,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV","json":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV.json","graph_json":"https://pith.science/api/pith-number/NLMEDV5DKMJQ7JA2VCX4HAR7GV/graph.json","events_json":"https://pith.science/api/pith-number/NLMEDV5DKMJQ7JA2VCX4HAR7GV/events.json","paper":"https://pith.science/paper/NLMEDV5D"},"agent_actions":{"view_html":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV","download_json":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV.json","view_paper":"https://pith.science/paper/NLMEDV5D","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.01525&json=true","fetch_graph":"https://pith.science/api/pith-number/NLMEDV5DKMJQ7JA2VCX4HAR7GV/graph.json","fetch_events":"https://pith.science/api/pith-number/NLMEDV5DKMJQ7JA2VCX4HAR7GV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV/action/storage_attestation","attest_author":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV/action/author_attestation","sign_citation":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV/action/citation_signature","submit_replication":"https://pith.science/pith/NLMEDV5DKMJQ7JA2VCX4HAR7GV/action/replication_record"}},"created_at":"2026-07-05T08:14:46.368060+00:00","updated_at":"2026-07-05T08:14:46.368060+00:00"}