{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JVW23KQH57JX2ZOS5OQP43AAC3","short_pith_number":"pith:JVW23KQH","schema_version":"1.0","canonical_sha256":"4d6dadaa07efd37d65d2eba0fe6c0016e604092f30bd4735f5174c27b4951e5f","source":{"kind":"arxiv","id":"2506.06877","version":2},"attestation_state":"computed","paper":{"title":"Right Is Not Enough: The Pitfalls of Outcome Supervision in Training LLMs for Math Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Da Zheng, Jiaxing Guo, Lun Du, Shengzhong Zhang, Tongshan Xu, Wenjie Yang, Zengfeng Huang","submitted_at":"2025-06-07T17:54:56Z","abstract_excerpt":"Outcome-rewarded Large Language Models (LLMs) have demonstrated remarkable success in mathematical problem-solving. However, this success often masks a critical issue: models frequently achieve correct answers through fundamentally unsound reasoning processes, a phenomenon indicative of reward hacking. We introduce MathOlympiadEval, a new dataset with fine-grained annotations, which reveals a significant gap between LLMs' answer correctness and their low process correctness. Existing automated methods like LLM-as-a-judge struggle to reliably detect these reasoning flaws. To address this, we pr"},"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":"2506.06877","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-06-07T17:54:56Z","cross_cats_sorted":[],"title_canon_sha256":"a4cbeb859acdbd571ba6ebf29e0f2754c4646e2e46132fdbee67f4c2de73193a","abstract_canon_sha256":"914040e8c9b01123e4d99e8593ce9b19e6147e4ed8dc8fa94a9477df98e6030c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:26:31.869535Z","signature_b64":"45t8ruWwt2lPsv8nW8WMUPvMwpL4Me5naTJchB8w67XJpeZS/xKgyPaf/DzE+eYymA8Aa/dgr2L9mBq0NRqVDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4d6dadaa07efd37d65d2eba0fe6c0016e604092f30bd4735f5174c27b4951e5f","last_reissued_at":"2026-07-05T11:26:31.869071Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:26:31.869071Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Right Is Not Enough: The Pitfalls of Outcome Supervision in Training LLMs for Math Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Da Zheng, Jiaxing Guo, Lun Du, Shengzhong Zhang, Tongshan Xu, Wenjie Yang, Zengfeng Huang","submitted_at":"2025-06-07T17:54:56Z","abstract_excerpt":"Outcome-rewarded Large Language Models (LLMs) have demonstrated remarkable success in mathematical problem-solving. However, this success often masks a critical issue: models frequently achieve correct answers through fundamentally unsound reasoning processes, a phenomenon indicative of reward hacking. We introduce MathOlympiadEval, a new dataset with fine-grained annotations, which reveals a significant gap between LLMs' answer correctness and their low process correctness. Existing automated methods like LLM-as-a-judge struggle to reliably detect these reasoning flaws. To address this, we pr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.06877","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/2506.06877/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":"2506.06877","created_at":"2026-07-05T11:26:31.869128+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.06877v2","created_at":"2026-07-05T11:26:31.869128+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.06877","created_at":"2026-07-05T11:26:31.869128+00:00"},{"alias_kind":"pith_short_12","alias_value":"JVW23KQH57JX","created_at":"2026-07-05T11:26:31.869128+00:00"},{"alias_kind":"pith_short_16","alias_value":"JVW23KQH57JX2ZOS","created_at":"2026-07-05T11:26:31.869128+00:00"},{"alias_kind":"pith_short_8","alias_value":"JVW23KQH","created_at":"2026-07-05T11:26:31.869128+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3","json":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3.json","graph_json":"https://pith.science/api/pith-number/JVW23KQH57JX2ZOS5OQP43AAC3/graph.json","events_json":"https://pith.science/api/pith-number/JVW23KQH57JX2ZOS5OQP43AAC3/events.json","paper":"https://pith.science/paper/JVW23KQH"},"agent_actions":{"view_html":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3","download_json":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3.json","view_paper":"https://pith.science/paper/JVW23KQH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.06877&json=true","fetch_graph":"https://pith.science/api/pith-number/JVW23KQH57JX2ZOS5OQP43AAC3/graph.json","fetch_events":"https://pith.science/api/pith-number/JVW23KQH57JX2ZOS5OQP43AAC3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3/action/storage_attestation","attest_author":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3/action/author_attestation","sign_citation":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3/action/citation_signature","submit_replication":"https://pith.science/pith/JVW23KQH57JX2ZOS5OQP43AAC3/action/replication_record"}},"created_at":"2026-07-05T11:26:31.869128+00:00","updated_at":"2026-07-05T11:26:31.869128+00:00"}