{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:PIFO2Z2SQJHT3W5Z7BLZY6WHHE","short_pith_number":"pith:PIFO2Z2S","schema_version":"1.0","canonical_sha256":"7a0aed6752824f3ddbb9f8579c7ac73929bd7a4eab9cfbdfcf2292ed0dcace5e","source":{"kind":"arxiv","id":"2507.09884","version":3},"attestation_state":"computed","paper":{"title":"VerifyBench: A Systematic Benchmark for Evaluating Reasoning Verifiers Across Domains","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Shiyu Hu, Wentao Zhang, Xuchen Li, Xuzhao Li, Yongzhen Guo","submitted_at":"2025-07-14T03:45:24Z","abstract_excerpt":"Large language models (LLMs) increasingly rely on reinforcement learning (RL) to enhance their reasoning capabilities through feedback. A critical challenge is verifying the consistency of model-generated responses and reference answers, since these responses are often lengthy, diverse, and nuanced. Rule-based verifiers struggle with complexity, prompting the use of model-based verifiers. However, specialized verifiers lack flexibility, while general LLM judges can be inconsistent. Existing research primarily focuses on building better verifiers, yet a systematic evaluation of different types "},"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":"2507.09884","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-07-14T03:45:24Z","cross_cats_sorted":[],"title_canon_sha256":"d34f58071fc1f87ab61a23d699cfd83c8c66fac604aa7a016937ebeb3a93315c","abstract_canon_sha256":"676ae3e81a4b0603b47a35d9e4a1ca6a8e4ab3a14efae500f17eb9289a98e3db"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:43:33.571976Z","signature_b64":"Dv+VtXsMWgkOrQYXdwltcgjOOrQ76AYfehlsvyn/7LANIlZQRPjdVN2t0nOrCcYYgHvW0dbld8bIBaG/jWBrBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7a0aed6752824f3ddbb9f8579c7ac73929bd7a4eab9cfbdfcf2292ed0dcace5e","last_reissued_at":"2026-07-05T11:43:33.571579Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:43:33.571579Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VerifyBench: A Systematic Benchmark for Evaluating Reasoning Verifiers Across Domains","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Shiyu Hu, Wentao Zhang, Xuchen Li, Xuzhao Li, Yongzhen Guo","submitted_at":"2025-07-14T03:45:24Z","abstract_excerpt":"Large language models (LLMs) increasingly rely on reinforcement learning (RL) to enhance their reasoning capabilities through feedback. A critical challenge is verifying the consistency of model-generated responses and reference answers, since these responses are often lengthy, diverse, and nuanced. Rule-based verifiers struggle with complexity, prompting the use of model-based verifiers. However, specialized verifiers lack flexibility, while general LLM judges can be inconsistent. Existing research primarily focuses on building better verifiers, yet a systematic evaluation of different types "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.09884","kind":"arxiv","version":3},"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/2507.09884/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":"2507.09884","created_at":"2026-07-05T11:43:33.571638+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.09884v3","created_at":"2026-07-05T11:43:33.571638+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.09884","created_at":"2026-07-05T11:43:33.571638+00:00"},{"alias_kind":"pith_short_12","alias_value":"PIFO2Z2SQJHT","created_at":"2026-07-05T11:43:33.571638+00:00"},{"alias_kind":"pith_short_16","alias_value":"PIFO2Z2SQJHT3W5Z","created_at":"2026-07-05T11:43:33.571638+00:00"},{"alias_kind":"pith_short_8","alias_value":"PIFO2Z2S","created_at":"2026-07-05T11:43:33.571638+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.01249","citing_title":"Trust Region On-Policy Distillation","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2510.00915","citing_title":"Reinforcement Learning with Verifiable yet Noisy Rewards under Imperfect Verifiers","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19196","citing_title":"Time to REFLECT: Can We Trust LLM Judges for Evidence-based Research Agents?","ref_index":25,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE","json":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE.json","graph_json":"https://pith.science/api/pith-number/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/graph.json","events_json":"https://pith.science/api/pith-number/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/events.json","paper":"https://pith.science/paper/PIFO2Z2S"},"agent_actions":{"view_html":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE","download_json":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE.json","view_paper":"https://pith.science/paper/PIFO2Z2S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.09884&json=true","fetch_graph":"https://pith.science/api/pith-number/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/graph.json","fetch_events":"https://pith.science/api/pith-number/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/action/storage_attestation","attest_author":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/action/author_attestation","sign_citation":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/action/citation_signature","submit_replication":"https://pith.science/pith/PIFO2Z2SQJHT3W5Z7BLZY6WHHE/action/replication_record"}},"created_at":"2026-07-05T11:43:33.571638+00:00","updated_at":"2026-07-05T11:43:33.571638+00:00"}