{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:NYHTCIAQSDUVVQI35GX5XU6S2A","short_pith_number":"pith:NYHTCIAQ","schema_version":"1.0","canonical_sha256":"6e0f31201090e95ac11be9afdbd3d2d03e07a094aae1efef9e31dceb9d070939","source":{"kind":"arxiv","id":"2311.02805","version":2},"attestation_state":"computed","paper":{"title":"Tailoring Self-Rationalizers with Multi-Reward Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aaron Chan, Brihi Joshi, Jack Hessel, Liunian Harold Li, Sahana Ramnath, Skyler Hallinan, Xiang Ren, Ximing Lu, Yejin Choi","submitted_at":"2023-11-06T00:20:11Z","abstract_excerpt":"Large language models (LMs) are capable of generating free-text rationales to aid question answering. However, prior work 1) suggests that useful self-rationalization is emergent only at significant scales (e.g., 175B parameter GPT-3); and 2) focuses largely on downstream performance, ignoring the semantics of the rationales themselves, e.g., are they faithful, true, and helpful for humans? In this work, we enable small-scale LMs (approx. 200x smaller than GPT-3) to generate rationales that not only improve downstream task performance, but are also more plausible, consistent, and diverse, asse"},"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":"2311.02805","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-06T00:20:11Z","cross_cats_sorted":[],"title_canon_sha256":"b18ba93c3c120e15394466e60d2b07c5bdb83f589ccee13ab376428262a6aeb9","abstract_canon_sha256":"d6a37fcd91cabeeef641ef852007f123c7acb1d210b2ed7666af2752503a04c3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:21:43.347641Z","signature_b64":"1PLIy6Dycanomg/1o1OiChlKljQXtNLwJOmQ5Vbr8sdSY89CdhaGTVVn2tBC0RY3qLYa6z0dn6UphGBMn3dpAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6e0f31201090e95ac11be9afdbd3d2d03e07a094aae1efef9e31dceb9d070939","last_reissued_at":"2026-07-05T08:21:43.347149Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:21:43.347149Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Tailoring Self-Rationalizers with Multi-Reward Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aaron Chan, Brihi Joshi, Jack Hessel, Liunian Harold Li, Sahana Ramnath, Skyler Hallinan, Xiang Ren, Ximing Lu, Yejin Choi","submitted_at":"2023-11-06T00:20:11Z","abstract_excerpt":"Large language models (LMs) are capable of generating free-text rationales to aid question answering. However, prior work 1) suggests that useful self-rationalization is emergent only at significant scales (e.g., 175B parameter GPT-3); and 2) focuses largely on downstream performance, ignoring the semantics of the rationales themselves, e.g., are they faithful, true, and helpful for humans? In this work, we enable small-scale LMs (approx. 200x smaller than GPT-3) to generate rationales that not only improve downstream task performance, but are also more plausible, consistent, and diverse, asse"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.02805","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/2311.02805/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":"2311.02805","created_at":"2026-07-05T08:21:43.347210+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.02805v2","created_at":"2026-07-05T08:21:43.347210+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.02805","created_at":"2026-07-05T08:21:43.347210+00:00"},{"alias_kind":"pith_short_12","alias_value":"NYHTCIAQSDUV","created_at":"2026-07-05T08:21:43.347210+00:00"},{"alias_kind":"pith_short_16","alias_value":"NYHTCIAQSDUVVQI3","created_at":"2026-07-05T08:21:43.347210+00:00"},{"alias_kind":"pith_short_8","alias_value":"NYHTCIAQ","created_at":"2026-07-05T08:21:43.347210+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A","json":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A.json","graph_json":"https://pith.science/api/pith-number/NYHTCIAQSDUVVQI35GX5XU6S2A/graph.json","events_json":"https://pith.science/api/pith-number/NYHTCIAQSDUVVQI35GX5XU6S2A/events.json","paper":"https://pith.science/paper/NYHTCIAQ"},"agent_actions":{"view_html":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A","download_json":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A.json","view_paper":"https://pith.science/paper/NYHTCIAQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.02805&json=true","fetch_graph":"https://pith.science/api/pith-number/NYHTCIAQSDUVVQI35GX5XU6S2A/graph.json","fetch_events":"https://pith.science/api/pith-number/NYHTCIAQSDUVVQI35GX5XU6S2A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A/action/storage_attestation","attest_author":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A/action/author_attestation","sign_citation":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A/action/citation_signature","submit_replication":"https://pith.science/pith/NYHTCIAQSDUVVQI35GX5XU6S2A/action/replication_record"}},"created_at":"2026-07-05T08:21:43.347210+00:00","updated_at":"2026-07-05T08:21:43.347210+00:00"}