{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:TLNZG57L5SUHPYIYXN7HEWKXZ2","short_pith_number":"pith:TLNZG57L","schema_version":"1.0","canonical_sha256":"9adb9377ebeca877e118bb7e725957ce881de4c7bf4fd904c1b98d2b27cab28c","source":{"kind":"arxiv","id":"2311.09766","version":4},"attestation_state":"computed","paper":{"title":"LLMs as Narcissistic Evaluators: When Ego Inflates Evaluation Scores","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chenghua Lin, Nafise Sadat Moosavi, Yiqi Liu","submitted_at":"2023-11-16T10:43:26Z","abstract_excerpt":"Automatic evaluation of generated textual content presents an ongoing challenge within the field of NLP. Given the impressive capabilities of modern language models (LMs) across diverse NLP tasks, there is a growing trend to employ these models in creating innovative evaluation metrics for automated assessment of generation tasks. This paper investigates a pivotal question: Do language model-driven evaluation metrics inherently exhibit bias favoring texts generated by the same underlying language model? Specifically, we assess whether prominent LM-based evaluation metrics (e.g. BARTScore, T5Sc"},"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":"2311.09766","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-16T10:43:26Z","cross_cats_sorted":[],"title_canon_sha256":"0eb38fe1dedf9c42ab3c44327668833144f65815f1282eadf83a2489c483a60c","abstract_canon_sha256":"41d49df5c64c2f7a38fc559209f8814eb4705abc2797d001a11d3f95ff73469d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:35.247598Z","signature_b64":"RlH0G7Ddgbbg6WxQay9T8/1Fswxz66y73fv5zS/y6nnEAqEZrujtfJfY9SLQD6im+SN+IVkk3G0IA6E+hQk6CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9adb9377ebeca877e118bb7e725957ce881de4c7bf4fd904c1b98d2b27cab28c","last_reissued_at":"2026-07-05T08:28:35.247102Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:35.247102Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLMs as Narcissistic Evaluators: When Ego Inflates Evaluation Scores","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chenghua Lin, Nafise Sadat Moosavi, Yiqi Liu","submitted_at":"2023-11-16T10:43:26Z","abstract_excerpt":"Automatic evaluation of generated textual content presents an ongoing challenge within the field of NLP. Given the impressive capabilities of modern language models (LMs) across diverse NLP tasks, there is a growing trend to employ these models in creating innovative evaluation metrics for automated assessment of generation tasks. This paper investigates a pivotal question: Do language model-driven evaluation metrics inherently exhibit bias favoring texts generated by the same underlying language model? Specifically, we assess whether prominent LM-based evaluation metrics (e.g. BARTScore, T5Sc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.09766","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/2311.09766/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":"2311.09766","created_at":"2026-07-05T08:28:35.247166+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.09766v4","created_at":"2026-07-05T08:28:35.247166+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.09766","created_at":"2026-07-05T08:28:35.247166+00:00"},{"alias_kind":"pith_short_12","alias_value":"TLNZG57L5SUH","created_at":"2026-07-05T08:28:35.247166+00:00"},{"alias_kind":"pith_short_16","alias_value":"TLNZG57L5SUHPYIY","created_at":"2026-07-05T08:28:35.247166+00:00"},{"alias_kind":"pith_short_8","alias_value":"TLNZG57L","created_at":"2026-07-05T08:28:35.247166+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05391","citing_title":"LLM-as-a-Verifier: A General-Purpose Verification Framework","ref_index":66,"is_internal_anchor":true},{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":208,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":207,"is_internal_anchor":false},{"citing_arxiv_id":"2408.15549","citing_title":"WildFeedback: Aligning LLMs With In-situ User Interactions And Feedback","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2404.13076","citing_title":"LLM Evaluators Recognize and Favor Their Own Generations","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10810","citing_title":"Likelihood scoring for continuations of mathematical text: a self-supervised benchmark with tests for shortcut vulnerabilities","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2504.07615","citing_title":"VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2","json":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2.json","graph_json":"https://pith.science/api/pith-number/TLNZG57L5SUHPYIYXN7HEWKXZ2/graph.json","events_json":"https://pith.science/api/pith-number/TLNZG57L5SUHPYIYXN7HEWKXZ2/events.json","paper":"https://pith.science/paper/TLNZG57L"},"agent_actions":{"view_html":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2","download_json":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2.json","view_paper":"https://pith.science/paper/TLNZG57L","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.09766&json=true","fetch_graph":"https://pith.science/api/pith-number/TLNZG57L5SUHPYIYXN7HEWKXZ2/graph.json","fetch_events":"https://pith.science/api/pith-number/TLNZG57L5SUHPYIYXN7HEWKXZ2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2/action/storage_attestation","attest_author":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2/action/author_attestation","sign_citation":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2/action/citation_signature","submit_replication":"https://pith.science/pith/TLNZG57L5SUHPYIYXN7HEWKXZ2/action/replication_record"}},"created_at":"2026-07-05T08:28:35.247166+00:00","updated_at":"2026-07-05T08:28:35.247166+00:00"}