{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6AESUCV3Y4TZPLW4FRNJOIK3YR","short_pith_number":"pith:6AESUCV3","schema_version":"1.0","canonical_sha256":"f0092a0abbc72797aedc2c5a97215bc46c3b69b4bb1e4093fc8b8ba7bdadb267","source":{"kind":"arxiv","id":"2606.09409","version":1},"attestation_state":"computed","paper":{"title":"Correct Looks Better: Pairwise Comparisons Reveal Accuracy Rankings","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Mina Remeli, Moritz Hardt","submitted_at":"2026-06-08T12:26:54Z","abstract_excerpt":"Pairwise comparisons combined with aggregation methods like Elo have become central to evaluating generative models, yet concerns remain that they reward superficial stylistic cues or display judge biases. In a more positive turn, we show that model rankings from pairwise comparisons strongly agree with ground-truth-based accuracy rankings when such ground truth is available for comparison. By converting five well-known benchmarks into free-form generative evaluations, we find that Elo rankings achieve a Spearman correlation above 0.9 with accuracy rankings and substantially outperform direct "},"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":"2606.09409","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-08T12:26:54Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"1cd378b574a3790c8589463a5027a113a4b81cdfe8162749e3974e087031094d","abstract_canon_sha256":"514e8570592994e1fc3843c241ce1ef8725664b8b04d4ddfb426ddf521e1e630"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T02:08:46.645561Z","signature_b64":"6d2twoYt9ZKGZ6zDTlDp9/0/iCEDuGWHOVy5xYNK9F4EJiBe/Gs5s+dS0C1ghVxGKkP4AGMl0EdAuDg5Ma8qDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f0092a0abbc72797aedc2c5a97215bc46c3b69b4bb1e4093fc8b8ba7bdadb267","last_reissued_at":"2026-06-09T02:08:46.644504Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T02:08:46.644504Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Correct Looks Better: Pairwise Comparisons Reveal Accuracy Rankings","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Mina Remeli, Moritz Hardt","submitted_at":"2026-06-08T12:26:54Z","abstract_excerpt":"Pairwise comparisons combined with aggregation methods like Elo have become central to evaluating generative models, yet concerns remain that they reward superficial stylistic cues or display judge biases. In a more positive turn, we show that model rankings from pairwise comparisons strongly agree with ground-truth-based accuracy rankings when such ground truth is available for comparison. By converting five well-known benchmarks into free-form generative evaluations, we find that Elo rankings achieve a Spearman correlation above 0.9 with accuracy rankings and substantially outperform direct "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.09409","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/2606.09409/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":"2606.09409","created_at":"2026-06-09T02:08:46.644648+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.09409v1","created_at":"2026-06-09T02:08:46.644648+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.09409","created_at":"2026-06-09T02:08:46.644648+00:00"},{"alias_kind":"pith_short_12","alias_value":"6AESUCV3Y4TZ","created_at":"2026-06-09T02:08:46.644648+00:00"},{"alias_kind":"pith_short_16","alias_value":"6AESUCV3Y4TZPLW4","created_at":"2026-06-09T02:08:46.644648+00:00"},{"alias_kind":"pith_short_8","alias_value":"6AESUCV3","created_at":"2026-06-09T02:08:46.644648+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/6AESUCV3Y4TZPLW4FRNJOIK3YR","json":"https://pith.science/pith/6AESUCV3Y4TZPLW4FRNJOIK3YR.json","graph_json":"https://pith.science/api/pith-number/6AESUCV3Y4TZPLW4FRNJOIK3YR/graph.json","events_json":"https://pith.science/api/pith-number/6AESUCV3Y4TZPLW4FRNJOIK3YR/events.json","paper":"https://pith.science/paper/6AESUCV3"},"agent_actions":{"view_html":"https://pith.science/pith/6AESUCV3Y4TZPLW4FRNJOIK3YR","download_json":"https://pith.science/pith/6AESUCV3Y4TZPLW4FRNJOIK3YR.json","view_paper":"https://pith.science/paper/6AESUCV3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.09409&json=true","fetch_graph":"https://pith.science/api/pith-number/6AESUCV3Y4TZPLW4FRNJOIK3YR/graph.json","fetch_events":"https://pith.science/api/pith-number/6AESUCV3Y4TZPLW4FRNJOIK3YR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6AESUCV3Y4TZPLW4FRNJOIK3YR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6AESUCV3Y4TZPLW4FRNJOIK3YR/action/storage_attestation","attest_author":"https://pith.science/pith/6AESUCV3Y4TZPLW4FRNJOIK3YR/action/author_attestation","sign_citation":"https://pith.science/pith/6AESUCV3Y4TZPLW4FRNJOIK3YR/action/citation_signature","submit_replication":"https://pith.science/pith/6AESUCV3Y4TZPLW4FRNJOIK3YR/action/replication_record"}},"created_at":"2026-06-09T02:08:46.644648+00:00","updated_at":"2026-06-09T02:08:46.644648+00:00"}