{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CMJIYLIEDJLKZINCXUBNY4ASYI","short_pith_number":"pith:CMJIYLIE","schema_version":"1.0","canonical_sha256":"13128c2d041a56aca1a2bd02dc7012c219e8661058eb34ab7ee933aac1385e1b","source":{"kind":"arxiv","id":"2507.10541","version":2},"attestation_state":"computed","paper":{"title":"REST: Stress Testing Large Reasoning Models by Asking Multiple Problems at Once","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Conghui He, H. Vicky Zhao, Lijun Wu, Qiyao Sun, Qizhi Pei, Yu Li, Zhuoshi Pan, Zinan Tang","submitted_at":"2025-07-14T17:58:47Z","abstract_excerpt":"Recent Large Reasoning Models (LRMs) have achieved remarkable progress on task-specific benchmarks, yet their evaluation methods remain constrained by isolated problem-solving paradigms. Existing benchmarks predominantly assess single-question reasoning through sequential testing, resulting critical limitations: (1) vulnerability to data contamination and less challenging (e.g., DeepSeek-R1 achieves 97.0% on MATH500), forcing costly creation of new questions with large human efforts, (2) failure to evaluate models under multi-context pressure, a key requirement for real-world deployment. To br"},"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":"2507.10541","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-14T17:58:47Z","cross_cats_sorted":[],"title_canon_sha256":"85bae915b00fb73fb66afee6262c0c933b68959212bc6b0fb0da69ad01a3b369","abstract_canon_sha256":"50b70a00d58b3567768dbd2180046bc2a7130c8e639cc9255f8fb98e8e7da727"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:37:26.894635Z","signature_b64":"MqmpSZ/9K+qR1gmS6hBchh70Qyf6LqGmRdPZv48zNRmJqewcYzEYSQD+uYQ99MB83eulZm6tT+fj5QH5nQzzBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"13128c2d041a56aca1a2bd02dc7012c219e8661058eb34ab7ee933aac1385e1b","last_reissued_at":"2026-07-05T11:37:26.894054Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:37:26.894054Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"REST: Stress Testing Large Reasoning Models by Asking Multiple Problems at Once","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Conghui He, H. Vicky Zhao, Lijun Wu, Qiyao Sun, Qizhi Pei, Yu Li, Zhuoshi Pan, Zinan Tang","submitted_at":"2025-07-14T17:58:47Z","abstract_excerpt":"Recent Large Reasoning Models (LRMs) have achieved remarkable progress on task-specific benchmarks, yet their evaluation methods remain constrained by isolated problem-solving paradigms. Existing benchmarks predominantly assess single-question reasoning through sequential testing, resulting critical limitations: (1) vulnerability to data contamination and less challenging (e.g., DeepSeek-R1 achieves 97.0% on MATH500), forcing costly creation of new questions with large human efforts, (2) failure to evaluate models under multi-context pressure, a key requirement for real-world deployment. To br"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.10541","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/2507.10541/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":"2507.10541","created_at":"2026-07-05T11:37:26.894113+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.10541v2","created_at":"2026-07-05T11:37:26.894113+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.10541","created_at":"2026-07-05T11:37:26.894113+00:00"},{"alias_kind":"pith_short_12","alias_value":"CMJIYLIEDJLK","created_at":"2026-07-05T11:37:26.894113+00:00"},{"alias_kind":"pith_short_16","alias_value":"CMJIYLIEDJLKZINC","created_at":"2026-07-05T11:37:26.894113+00:00"},{"alias_kind":"pith_short_8","alias_value":"CMJIYLIE","created_at":"2026-07-05T11:37:26.894113+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07321","citing_title":"From Atomic Actions to Standard Operating Procedures: Iterative Tool Optimization for Self-Evolving LLM Agents","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2604.10480","citing_title":"Tracing the Roots: A Multi-Agent Framework for Uncovering Data Lineage in Post-Training LLMs","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI","json":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI.json","graph_json":"https://pith.science/api/pith-number/CMJIYLIEDJLKZINCXUBNY4ASYI/graph.json","events_json":"https://pith.science/api/pith-number/CMJIYLIEDJLKZINCXUBNY4ASYI/events.json","paper":"https://pith.science/paper/CMJIYLIE"},"agent_actions":{"view_html":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI","download_json":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI.json","view_paper":"https://pith.science/paper/CMJIYLIE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.10541&json=true","fetch_graph":"https://pith.science/api/pith-number/CMJIYLIEDJLKZINCXUBNY4ASYI/graph.json","fetch_events":"https://pith.science/api/pith-number/CMJIYLIEDJLKZINCXUBNY4ASYI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI/action/storage_attestation","attest_author":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI/action/author_attestation","sign_citation":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI/action/citation_signature","submit_replication":"https://pith.science/pith/CMJIYLIEDJLKZINCXUBNY4ASYI/action/replication_record"}},"created_at":"2026-07-05T11:37:26.894113+00:00","updated_at":"2026-07-05T11:37:26.894113+00:00"}