{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:EGPQISGQSIEUUWDKLFUEKGZXCI","short_pith_number":"pith:EGPQISGQ","schema_version":"1.0","canonical_sha256":"219f0448d092094a586a5968451b37122f9f6d9850f090a3e5330f64046d9799","source":{"kind":"arxiv","id":"2504.15253","version":2},"attestation_state":"computed","paper":{"title":"Evaluating Judges as Evaluators: The JETTS Benchmark of LLM-as-Judges as Test-Time Scaling Evaluators","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Austin Xu, Caiming Xiong, Peifeng Wang, Shafiq Joty, Yilun Zhou","submitted_at":"2025-04-21T17:33:23Z","abstract_excerpt":"Scaling test-time computation, or affording a generator large language model (LLM) extra compute during inference, typically employs the help of external non-generative evaluators (i.e., reward models). Concurrently, LLM-judges, models trained to generate evaluations and critiques (explanations) in natural language, are becoming increasingly popular in automatic evaluation. Despite judge empirical successes, their effectiveness as evaluators in test-time scaling settings is largely unknown. In this paper, we introduce the Judge Evaluation for Test-Time Scaling (JETTS) benchmark, which evaluate"},"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":"2504.15253","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-21T17:33:23Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"588ff1f300b4f69de07667a02b5126fbe2d306c7a040cad27c1bcd88e3c9df17","abstract_canon_sha256":"be7b7877e59aa199eb52028c32047114fe98655b16e74d032ccfec2fb1344a98"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:07:23.882176Z","signature_b64":"kh/UWsEJx4JTxRdyUb5RWDFSyLqqJ3KYSm+otvY8cCSY9UTI45Jwzde5uGG4ZfdTEPFUhxy30qDhqE6wEUQGCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"219f0448d092094a586a5968451b37122f9f6d9850f090a3e5330f64046d9799","last_reissued_at":"2026-07-05T11:07:23.881696Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:07:23.881696Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evaluating Judges as Evaluators: The JETTS Benchmark of LLM-as-Judges as Test-Time Scaling Evaluators","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Austin Xu, Caiming Xiong, Peifeng Wang, Shafiq Joty, Yilun Zhou","submitted_at":"2025-04-21T17:33:23Z","abstract_excerpt":"Scaling test-time computation, or affording a generator large language model (LLM) extra compute during inference, typically employs the help of external non-generative evaluators (i.e., reward models). Concurrently, LLM-judges, models trained to generate evaluations and critiques (explanations) in natural language, are becoming increasingly popular in automatic evaluation. Despite judge empirical successes, their effectiveness as evaluators in test-time scaling settings is largely unknown. In this paper, we introduce the Judge Evaluation for Test-Time Scaling (JETTS) benchmark, which evaluate"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.15253","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/2504.15253/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":"2504.15253","created_at":"2026-07-05T11:07:23.881750+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.15253v2","created_at":"2026-07-05T11:07:23.881750+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.15253","created_at":"2026-07-05T11:07:23.881750+00:00"},{"alias_kind":"pith_short_12","alias_value":"EGPQISGQSIEU","created_at":"2026-07-05T11:07:23.881750+00:00"},{"alias_kind":"pith_short_16","alias_value":"EGPQISGQSIEUUWDK","created_at":"2026-07-05T11:07:23.881750+00:00"},{"alias_kind":"pith_short_8","alias_value":"EGPQISGQ","created_at":"2026-07-05T11:07:23.881750+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21627","citing_title":"Counsel: A Meta-Evaluation Dataset for Agentic Tasks","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05384","citing_title":"Stability vs. Manipulability: Evaluating Robustness Under Post-Decision Interaction in LLM Judges","ref_index":81,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01462","citing_title":"An Enigma of Artificial Reason: Investigating the Production-Evaluation Gap in Large Reasoning Models","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2509.23542","citing_title":"On the Shelf Life of Fine-Tuned LLM-Judges: Future-Proofing, Backward-Compatibility, and Question Generalization","ref_index":54,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI","json":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI.json","graph_json":"https://pith.science/api/pith-number/EGPQISGQSIEUUWDKLFUEKGZXCI/graph.json","events_json":"https://pith.science/api/pith-number/EGPQISGQSIEUUWDKLFUEKGZXCI/events.json","paper":"https://pith.science/paper/EGPQISGQ"},"agent_actions":{"view_html":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI","download_json":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI.json","view_paper":"https://pith.science/paper/EGPQISGQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.15253&json=true","fetch_graph":"https://pith.science/api/pith-number/EGPQISGQSIEUUWDKLFUEKGZXCI/graph.json","fetch_events":"https://pith.science/api/pith-number/EGPQISGQSIEUUWDKLFUEKGZXCI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI/action/storage_attestation","attest_author":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI/action/author_attestation","sign_citation":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI/action/citation_signature","submit_replication":"https://pith.science/pith/EGPQISGQSIEUUWDKLFUEKGZXCI/action/replication_record"}},"created_at":"2026-07-05T11:07:23.881750+00:00","updated_at":"2026-07-05T11:07:23.881750+00:00"}