{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BOFNM75YPDHYJTA5SDQZKWEXTC","short_pith_number":"pith:BOFNM75Y","schema_version":"1.0","canonical_sha256":"0b8ad67fb878cf84cc1d90e1955897989999399bf31b9fd8107810dd291b94a8","source":{"kind":"arxiv","id":"2503.04381","version":2},"attestation_state":"computed","paper":{"title":"TRACT: Regression-Aware Fine-tuning Meets Chain-of-Thought Reasoning for LLM-as-a-Judge","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Cheng-Han Chiang, Hung-yi Lee, Michal Lukasik","submitted_at":"2025-03-06T12:33:20Z","abstract_excerpt":"The LLM-as-a-judge paradigm uses large language models (LLMs) for automated text evaluation, where a numerical assessment is assigned by an LLM to the input text following scoring rubrics. Existing methods for LLM-as-a-judge use cross-entropy (CE) loss for fine-tuning, which neglects the numeric nature of score prediction. Recent work addresses numerical prediction limitations of LLM fine-tuning through regression-aware fine-tuning, which, however, does not consider chain-of-thought (CoT) reasoning for score prediction. In this paper, we introduce TRACT (Two-stage Regression-Aware fine-tuning "},"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":"2503.04381","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-06T12:33:20Z","cross_cats_sorted":[],"title_canon_sha256":"d1293ef5b1dd7043c7c3041c004e77c1150ee545b49918b0c5555c6d94928528","abstract_canon_sha256":"3d8848b9fef45082c85eb3463a51d56b01467d83deb3cf84ad801a8ec8566409"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:56.393141Z","signature_b64":"G+M1XtszGn9ys3e7pIBVhDLp5dWhZ6pNvYPbnNSuPITGVgXVcCkKZI2UODfSClX4rWu7/WdGg68iLYpkXLhyCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0b8ad67fb878cf84cc1d90e1955897989999399bf31b9fd8107810dd291b94a8","last_reissued_at":"2026-07-05T11:16:56.392646Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:56.392646Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TRACT: Regression-Aware Fine-tuning Meets Chain-of-Thought Reasoning for LLM-as-a-Judge","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Cheng-Han Chiang, Hung-yi Lee, Michal Lukasik","submitted_at":"2025-03-06T12:33:20Z","abstract_excerpt":"The LLM-as-a-judge paradigm uses large language models (LLMs) for automated text evaluation, where a numerical assessment is assigned by an LLM to the input text following scoring rubrics. Existing methods for LLM-as-a-judge use cross-entropy (CE) loss for fine-tuning, which neglects the numeric nature of score prediction. Recent work addresses numerical prediction limitations of LLM fine-tuning through regression-aware fine-tuning, which, however, does not consider chain-of-thought (CoT) reasoning for score prediction. In this paper, we introduce TRACT (Two-stage Regression-Aware fine-tuning "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.04381","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/2503.04381/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":"2503.04381","created_at":"2026-07-05T11:16:56.392707+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.04381v2","created_at":"2026-07-05T11:16:56.392707+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.04381","created_at":"2026-07-05T11:16:56.392707+00:00"},{"alias_kind":"pith_short_12","alias_value":"BOFNM75YPDHY","created_at":"2026-07-05T11:16:56.392707+00:00"},{"alias_kind":"pith_short_16","alias_value":"BOFNM75YPDHYJTA5","created_at":"2026-07-05T11:16:56.392707+00:00"},{"alias_kind":"pith_short_8","alias_value":"BOFNM75Y","created_at":"2026-07-05T11:16:56.392707+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.08147","citing_title":"Biological Reasoning-Informed Regression for Interpretable Regulatory DNA Activity Prediction","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2504.02181","citing_title":"A Survey of Scaling in Large Language Model Reasoning","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC","json":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC.json","graph_json":"https://pith.science/api/pith-number/BOFNM75YPDHYJTA5SDQZKWEXTC/graph.json","events_json":"https://pith.science/api/pith-number/BOFNM75YPDHYJTA5SDQZKWEXTC/events.json","paper":"https://pith.science/paper/BOFNM75Y"},"agent_actions":{"view_html":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC","download_json":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC.json","view_paper":"https://pith.science/paper/BOFNM75Y","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.04381&json=true","fetch_graph":"https://pith.science/api/pith-number/BOFNM75YPDHYJTA5SDQZKWEXTC/graph.json","fetch_events":"https://pith.science/api/pith-number/BOFNM75YPDHYJTA5SDQZKWEXTC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC/action/storage_attestation","attest_author":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC/action/author_attestation","sign_citation":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC/action/citation_signature","submit_replication":"https://pith.science/pith/BOFNM75YPDHYJTA5SDQZKWEXTC/action/replication_record"}},"created_at":"2026-07-05T11:16:56.392707+00:00","updated_at":"2026-07-05T11:16:56.392707+00:00"}