{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UE47YYAL6OFTPJ2DBUISYRO5HK","short_pith_number":"pith:UE47YYAL","schema_version":"1.0","canonical_sha256":"a139fc600bf38b37a7430d112c45dd3ab825151ed0667da82333433d6f9ce2df","source":{"kind":"arxiv","id":"2601.22548","version":4},"attestation_state":"computed","paper":{"title":"Are LLM Evaluators Really Narcissists? Sanity Checking Self-Preference Evaluations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Dani Roytburg, Jou Barzdukas, Mackenzie Puig-Hall, Matthew Bozoukov, Matthew Nguyen, Narmeen Oozeer","submitted_at":"2026-01-30T04:38:18Z","abstract_excerpt":"Recent research has shown that large language models (LLMs) favor their own outputs when acting as judges, undermining the integrity of automated post-training and evaluation workflows. However, it is difficult to disentangle which behaviors are explained by narcissism versus experimental confounds. Specifically, LLM evaluators may deliver self-preferring verdicts when comparing responses to questions they fail on; these verdicts may not depend on the identity of the author, but on evaluator quality. We correct this by directly comparing the judge's voting distribution in cases where it evalua"},"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":"2601.22548","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-30T04:38:18Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"c7146fe3ac3763d080d19234f945dd02fe6d1d265a0c338466d051e5323246e2","abstract_canon_sha256":"4ddfc97ec4f37f263ec4a98c5b8003a8316d2c83bcb2438fc5fcfdbd01fba98d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-24T00:14:22.354394Z","signature_b64":"cJQnLCoBqJgUVhz+1ZZjs1usP31YlMGGqiP/87cstLXZkTATfKNzOcTIv63/mN2ZbmiYFT78X+uxOarZCE4cDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a139fc600bf38b37a7430d112c45dd3ab825151ed0667da82333433d6f9ce2df","last_reissued_at":"2026-06-24T00:14:22.353959Z","signature_status":"signed_v1","first_computed_at":"2026-06-24T00:14:22.353959Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Are LLM Evaluators Really Narcissists? Sanity Checking Self-Preference Evaluations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Dani Roytburg, Jou Barzdukas, Mackenzie Puig-Hall, Matthew Bozoukov, Matthew Nguyen, Narmeen Oozeer","submitted_at":"2026-01-30T04:38:18Z","abstract_excerpt":"Recent research has shown that large language models (LLMs) favor their own outputs when acting as judges, undermining the integrity of automated post-training and evaluation workflows. However, it is difficult to disentangle which behaviors are explained by narcissism versus experimental confounds. Specifically, LLM evaluators may deliver self-preferring verdicts when comparing responses to questions they fail on; these verdicts may not depend on the identity of the author, but on evaluator quality. We correct this by directly comparing the judge's voting distribution in cases where it evalua"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.22548","kind":"arxiv","version":4},"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/2601.22548/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":"2601.22548","created_at":"2026-06-24T00:14:22.354015+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.22548v4","created_at":"2026-06-24T00:14:22.354015+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.22548","created_at":"2026-06-24T00:14:22.354015+00:00"},{"alias_kind":"pith_short_12","alias_value":"UE47YYAL6OFT","created_at":"2026-06-24T00:14:22.354015+00:00"},{"alias_kind":"pith_short_16","alias_value":"UE47YYAL6OFTPJ2D","created_at":"2026-06-24T00:14:22.354015+00:00"},{"alias_kind":"pith_short_8","alias_value":"UE47YYAL","created_at":"2026-06-24T00:14:22.354015+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/UE47YYAL6OFTPJ2DBUISYRO5HK","json":"https://pith.science/pith/UE47YYAL6OFTPJ2DBUISYRO5HK.json","graph_json":"https://pith.science/api/pith-number/UE47YYAL6OFTPJ2DBUISYRO5HK/graph.json","events_json":"https://pith.science/api/pith-number/UE47YYAL6OFTPJ2DBUISYRO5HK/events.json","paper":"https://pith.science/paper/UE47YYAL"},"agent_actions":{"view_html":"https://pith.science/pith/UE47YYAL6OFTPJ2DBUISYRO5HK","download_json":"https://pith.science/pith/UE47YYAL6OFTPJ2DBUISYRO5HK.json","view_paper":"https://pith.science/paper/UE47YYAL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.22548&json=true","fetch_graph":"https://pith.science/api/pith-number/UE47YYAL6OFTPJ2DBUISYRO5HK/graph.json","fetch_events":"https://pith.science/api/pith-number/UE47YYAL6OFTPJ2DBUISYRO5HK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UE47YYAL6OFTPJ2DBUISYRO5HK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UE47YYAL6OFTPJ2DBUISYRO5HK/action/storage_attestation","attest_author":"https://pith.science/pith/UE47YYAL6OFTPJ2DBUISYRO5HK/action/author_attestation","sign_citation":"https://pith.science/pith/UE47YYAL6OFTPJ2DBUISYRO5HK/action/citation_signature","submit_replication":"https://pith.science/pith/UE47YYAL6OFTPJ2DBUISYRO5HK/action/replication_record"}},"created_at":"2026-06-24T00:14:22.354015+00:00","updated_at":"2026-06-24T00:14:22.354015+00:00"}