{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:5O7LBIGX76TAQWGGXTIE6JLOIP","short_pith_number":"pith:5O7LBIGX","schema_version":"1.0","canonical_sha256":"ebbeb0a0d7ffa60858c6bcd04f256e43e6331674a9aa47e9afe935d7102b4931","source":{"kind":"arxiv","id":"2207.00779","version":2},"attestation_state":"computed","paper":{"title":"FRAME: Evaluating Rationale-Label Consistency Metrics for Free-Text Rationales","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aaron Chan, Hamed Firooz, Liang Tan, Maziar Sanjabi, Shaoliang Nie, Xiang Ren, Xiaochang Peng","submitted_at":"2022-07-02T09:25:29Z","abstract_excerpt":"Following how humans communicate, free-text rationales aim to use natural language to explain neural language model (LM) behavior. However, free-text rationales' unconstrained nature makes them prone to hallucination, so it is important to have metrics for free-text rationale quality. Existing free-text rationale metrics measure how consistent the rationale is with the LM's predicted label, but there is no protocol for assessing such metrics' reliability. Thus, we propose FRAME, a framework for evaluating rationale-label consistency (RLC) metrics for free-text rationales. FRAME is based on thr"},"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":"2207.00779","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-07-02T09:25:29Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"a081a0827a82a49c341d8dfd633b3f93b5972867092ffcaf52e7c09076b24941","abstract_canon_sha256":"29b437d66645c2bc5bfaa5ca982c882eec21243933b35e8847566d7e9c9e067e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:26:44.425344Z","signature_b64":"zpklJnMMuRHvV3qesULV8jCEBOoGSaBhfHH8xThClUEsNyQeulBjo+3FxC4bl/+diHKVYn8fd+N3obIyR8OfBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ebbeb0a0d7ffa60858c6bcd04f256e43e6331674a9aa47e9afe935d7102b4931","last_reissued_at":"2026-07-05T05:26:44.424916Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:26:44.424916Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FRAME: Evaluating Rationale-Label Consistency Metrics for Free-Text Rationales","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aaron Chan, Hamed Firooz, Liang Tan, Maziar Sanjabi, Shaoliang Nie, Xiang Ren, Xiaochang Peng","submitted_at":"2022-07-02T09:25:29Z","abstract_excerpt":"Following how humans communicate, free-text rationales aim to use natural language to explain neural language model (LM) behavior. However, free-text rationales' unconstrained nature makes them prone to hallucination, so it is important to have metrics for free-text rationale quality. Existing free-text rationale metrics measure how consistent the rationale is with the LM's predicted label, but there is no protocol for assessing such metrics' reliability. Thus, we propose FRAME, a framework for evaluating rationale-label consistency (RLC) metrics for free-text rationales. FRAME is based on thr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2207.00779","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/2207.00779/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":"2207.00779","created_at":"2026-07-05T05:26:44.424970+00:00"},{"alias_kind":"arxiv_version","alias_value":"2207.00779v2","created_at":"2026-07-05T05:26:44.424970+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2207.00779","created_at":"2026-07-05T05:26:44.424970+00:00"},{"alias_kind":"pith_short_12","alias_value":"5O7LBIGX76TA","created_at":"2026-07-05T05:26:44.424970+00:00"},{"alias_kind":"pith_short_16","alias_value":"5O7LBIGX76TAQWGG","created_at":"2026-07-05T05:26:44.424970+00:00"},{"alias_kind":"pith_short_8","alias_value":"5O7LBIGX","created_at":"2026-07-05T05:26:44.424970+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.12169","citing_title":"Regulation of Language Models With Interpretability Will Likely Result In A Performance Trade-Off","ref_index":9,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP","json":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP.json","graph_json":"https://pith.science/api/pith-number/5O7LBIGX76TAQWGGXTIE6JLOIP/graph.json","events_json":"https://pith.science/api/pith-number/5O7LBIGX76TAQWGGXTIE6JLOIP/events.json","paper":"https://pith.science/paper/5O7LBIGX"},"agent_actions":{"view_html":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP","download_json":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP.json","view_paper":"https://pith.science/paper/5O7LBIGX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2207.00779&json=true","fetch_graph":"https://pith.science/api/pith-number/5O7LBIGX76TAQWGGXTIE6JLOIP/graph.json","fetch_events":"https://pith.science/api/pith-number/5O7LBIGX76TAQWGGXTIE6JLOIP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP/action/storage_attestation","attest_author":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP/action/author_attestation","sign_citation":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP/action/citation_signature","submit_replication":"https://pith.science/pith/5O7LBIGX76TAQWGGXTIE6JLOIP/action/replication_record"}},"created_at":"2026-07-05T05:26:44.424970+00:00","updated_at":"2026-07-05T05:26:44.424970+00:00"}