{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3VZ6LFOZMHCMMNR4QEJ2D4UU3A","short_pith_number":"pith:3VZ6LFOZ","schema_version":"1.0","canonical_sha256":"dd73e595d961c4c6363c8113a1f294d80d4f9cda424c46c5b9a47d0f6a502bed","source":{"kind":"arxiv","id":"2302.14520","version":2},"attestation_state":"computed","paper":{"title":"Large Language Models Are State-of-the-Art Evaluators of Translation Quality","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Christian Federmann, Tom Kocmi","submitted_at":"2023-02-28T12:23:48Z","abstract_excerpt":"We describe GEMBA, a GPT-based metric for assessment of translation quality, which works both with a reference translation and without. In our evaluation, we focus on zero-shot prompting, comparing four prompt variants in two modes, based on the availability of the reference. We investigate nine versions of GPT models, including ChatGPT and GPT-4. We show that our method for translation quality assessment only works with GPT~3.5 and larger models. Comparing to results from WMT22's Metrics shared task, our method achieves state-of-the-art accuracy in both modes when compared to MQM-based human "},"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":"2302.14520","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-28T12:23:48Z","cross_cats_sorted":[],"title_canon_sha256":"fa9f3cd114f94c24290e1120c6881c9ee976e94a7a2bdf77d87e4ae7c0c2daed","abstract_canon_sha256":"cb091eb9c738c663dad2b12c8b797264057792a82dd56bbeb86d874e95525031"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:16:16.657117Z","signature_b64":"KhZGlMvLBNr4odTA1xcM7SWtQKEKBc0B8chdOkRA1R1/YfO/4IJ8FD0Caa4Juh7nFMw3BeBPPlKbdcluiLD3AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dd73e595d961c4c6363c8113a1f294d80d4f9cda424c46c5b9a47d0f6a502bed","last_reissued_at":"2026-07-05T06:16:16.656618Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:16:16.656618Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Large Language Models Are State-of-the-Art Evaluators of Translation Quality","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Christian Federmann, Tom Kocmi","submitted_at":"2023-02-28T12:23:48Z","abstract_excerpt":"We describe GEMBA, a GPT-based metric for assessment of translation quality, which works both with a reference translation and without. In our evaluation, we focus on zero-shot prompting, comparing four prompt variants in two modes, based on the availability of the reference. We investigate nine versions of GPT models, including ChatGPT and GPT-4. We show that our method for translation quality assessment only works with GPT~3.5 and larger models. Comparing to results from WMT22's Metrics shared task, our method achieves state-of-the-art accuracy in both modes when compared to MQM-based human "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.14520","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/2302.14520/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":"2302.14520","created_at":"2026-07-05T06:16:16.656674+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.14520v2","created_at":"2026-07-05T06:16:16.656674+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.14520","created_at":"2026-07-05T06:16:16.656674+00:00"},{"alias_kind":"pith_short_12","alias_value":"3VZ6LFOZMHCM","created_at":"2026-07-05T06:16:16.656674+00:00"},{"alias_kind":"pith_short_16","alias_value":"3VZ6LFOZMHCMMNR4","created_at":"2026-07-05T06:16:16.656674+00:00"},{"alias_kind":"pith_short_8","alias_value":"3VZ6LFOZ","created_at":"2026-07-05T06:16:16.656674+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07989","citing_title":"Who Broke the System? Failure Localization in LLM-Based Multi-Agent Systems","ref_index":73,"is_internal_anchor":true},{"citing_arxiv_id":"2605.29430","citing_title":"Towards Human-Like Interactive Speech Recognition With Agentic Correction and Semantic Evaluation","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.31080","citing_title":"A Pilot Study on Curator-Guided Multilingual Art Description for Blind and Low-Vision Audiences with Small Vision-Language Models","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01136","citing_title":"From Outliers to Errors: Auditing Pali-to-English LLM Translations with Multi-Reference Adjudication","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17041","citing_title":"Agentic AI Translate: An Agentic Translator Prototype for Translation as Communication Design","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2404.18796","citing_title":"Replacing Judges with Juries: Evaluating LLM Generations with a Panel of Diverse Models","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20334","citing_title":"An Explainable Approach to Document-level Translation Evaluation with Topic Modeling","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08851","citing_title":"Cross-Lingual Attention Distillation with Personality-Informed Generative Augmentation for Multilingual Personality Recognition","ref_index":30,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A","json":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A.json","graph_json":"https://pith.science/api/pith-number/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/graph.json","events_json":"https://pith.science/api/pith-number/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/events.json","paper":"https://pith.science/paper/3VZ6LFOZ"},"agent_actions":{"view_html":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A","download_json":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A.json","view_paper":"https://pith.science/paper/3VZ6LFOZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.14520&json=true","fetch_graph":"https://pith.science/api/pith-number/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/graph.json","fetch_events":"https://pith.science/api/pith-number/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/action/storage_attestation","attest_author":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/action/author_attestation","sign_citation":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/action/citation_signature","submit_replication":"https://pith.science/pith/3VZ6LFOZMHCMMNR4QEJ2D4UU3A/action/replication_record"}},"created_at":"2026-07-05T06:16:16.656674+00:00","updated_at":"2026-07-05T06:16:16.656674+00:00"}