{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:XDNFZRDQJWOTVQEUWPJX7RF6LB","short_pith_number":"pith:XDNFZRDQ","canonical_record":{"source":{"id":"2407.08733","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T17:58:58Z","cross_cats_sorted":[],"title_canon_sha256":"97e9905fe81b509c44b589ce53c43f1f9f28634eb10e3375782c80a2f09e32c1","abstract_canon_sha256":"b325b5276d4ac0f8a0dc08987f0e8db27cd8bedb37e74a02fdd259999f84f62e"},"schema_version":"1.0"},"canonical_sha256":"b8da5cc4704d9d3ac094b3d37fc4be584a1602c976b0a623b8f04a8729c89249","source":{"kind":"arxiv","id":"2407.08733","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.08733","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"arxiv_version","alias_value":"2407.08733v2","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.08733","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"pith_short_12","alias_value":"XDNFZRDQJWOT","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"pith_short_16","alias_value":"XDNFZRDQJWOTVQEU","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"pith_short_8","alias_value":"XDNFZRDQ","created_at":"2026-07-05T09:17:45Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:XDNFZRDQJWOTVQEUWPJX7RF6LB","target":"record","payload":{"canonical_record":{"source":{"id":"2407.08733","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T17:58:58Z","cross_cats_sorted":[],"title_canon_sha256":"97e9905fe81b509c44b589ce53c43f1f9f28634eb10e3375782c80a2f09e32c1","abstract_canon_sha256":"b325b5276d4ac0f8a0dc08987f0e8db27cd8bedb37e74a02fdd259999f84f62e"},"schema_version":"1.0"},"canonical_sha256":"b8da5cc4704d9d3ac094b3d37fc4be584a1602c976b0a623b8f04a8729c89249","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:17:45.829557Z","signature_b64":"pTSYvGye2QGiqh+Sl9Yzg2ny32BdRqZTbRbLvw8py76wPf2on7Mj3XmnScNgAVA7OkBlABZGfE68oAzLW4+KAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b8da5cc4704d9d3ac094b3d37fc4be584a1602c976b0a623b8f04a8729c89249","last_reissued_at":"2026-07-05T09:17:45.829095Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:17:45.829095Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2407.08733","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:17:45Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uP735Z6EQ2+zLHvZLHiqLsIX8ZtoI0QAyUzS+zu9ZMeI35WfTzhO203G/icL2KFa+NBZdXW1YOL2p9IJsyQbDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T20:52:03.508522Z"},"content_sha256":"64f396a7b4c21c75a5b11b672cca217ec2b1f6b9473ecab2cd3687ef0e738e3a","schema_version":"1.0","event_id":"sha256:64f396a7b4c21c75a5b11b672cca217ec2b1f6b9473ecab2cd3687ef0e738e3a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:XDNFZRDQJWOTVQEUWPJX7RF6LB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Is Your Model Really A Good Math Reasoner? Evaluating Mathematical Reasoning with Checklist","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Derek F. Wong, Jindong Wang, Kaizhu Huang, Maizhen Ning, Qiufeng Wang, Shudong Liu, Wei Liu, Xiaowei Huang, Zihao Zhou","submitted_at":"2024-07-11T17:58:58Z","abstract_excerpt":"Exceptional mathematical reasoning ability is one of the key features that demonstrate the power of large language models (LLMs). How to comprehensively define and evaluate the mathematical abilities of LLMs, and even reflect the user experience in real-world scenarios, has emerged as a critical issue. Current benchmarks predominantly concentrate on problem-solving capabilities, presenting a substantial risk of model overfitting and fails to accurately measure the genuine mathematical reasoning abilities. In this paper, we argue that if a model really understands a problem, it should be robust"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.08733","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/2407.08733/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:17:45Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qAvTRgr+UJKiHLDdCauJ7J9CkhnIZRftDc8OZTHjiucYjMOFcwjh/GE0mU39jrbglWy9eQlT5zlZrbDIEWVaCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T20:52:03.509435Z"},"content_sha256":"b84595f2d089d48843acf57031f71c30441c209658d94328bdcd37e4c01a4c4d","schema_version":"1.0","event_id":"sha256:b84595f2d089d48843acf57031f71c30441c209658d94328bdcd37e4c01a4c4d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/XDNFZRDQJWOTVQEUWPJX7RF6LB/bundle.json","state_url":"https://pith.science/pith/XDNFZRDQJWOTVQEUWPJX7RF6LB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/XDNFZRDQJWOTVQEUWPJX7RF6LB/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-09T20:52:03Z","links":{"resolver":"https://pith.science/pith/XDNFZRDQJWOTVQEUWPJX7RF6LB","bundle":"https://pith.science/pith/XDNFZRDQJWOTVQEUWPJX7RF6LB/bundle.json","state":"https://pith.science/pith/XDNFZRDQJWOTVQEUWPJX7RF6LB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/XDNFZRDQJWOTVQEUWPJX7RF6LB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:XDNFZRDQJWOTVQEUWPJX7RF6LB","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b325b5276d4ac0f8a0dc08987f0e8db27cd8bedb37e74a02fdd259999f84f62e","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T17:58:58Z","title_canon_sha256":"97e9905fe81b509c44b589ce53c43f1f9f28634eb10e3375782c80a2f09e32c1"},"schema_version":"1.0","source":{"id":"2407.08733","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.08733","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"arxiv_version","alias_value":"2407.08733v2","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.08733","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"pith_short_12","alias_value":"XDNFZRDQJWOT","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"pith_short_16","alias_value":"XDNFZRDQJWOTVQEU","created_at":"2026-07-05T09:17:45Z"},{"alias_kind":"pith_short_8","alias_value":"XDNFZRDQ","created_at":"2026-07-05T09:17:45Z"}],"graph_snapshots":[{"event_id":"sha256:b84595f2d089d48843acf57031f71c30441c209658d94328bdcd37e4c01a4c4d","target":"graph","created_at":"2026-07-05T09:17:45Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2407.08733/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Exceptional mathematical reasoning ability is one of the key features that demonstrate the power of large language models (LLMs). How to comprehensively define and evaluate the mathematical abilities of LLMs, and even reflect the user experience in real-world scenarios, has emerged as a critical issue. Current benchmarks predominantly concentrate on problem-solving capabilities, presenting a substantial risk of model overfitting and fails to accurately measure the genuine mathematical reasoning abilities. In this paper, we argue that if a model really understands a problem, it should be robust","authors_text":"Derek F. Wong, Jindong Wang, Kaizhu Huang, Maizhen Ning, Qiufeng Wang, Shudong Liu, Wei Liu, Xiaowei Huang, Zihao Zhou","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T17:58:58Z","title":"Is Your Model Really A Good Math Reasoner? Evaluating Mathematical Reasoning with Checklist"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.08733","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:64f396a7b4c21c75a5b11b672cca217ec2b1f6b9473ecab2cd3687ef0e738e3a","target":"record","created_at":"2026-07-05T09:17:45Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b325b5276d4ac0f8a0dc08987f0e8db27cd8bedb37e74a02fdd259999f84f62e","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-11T17:58:58Z","title_canon_sha256":"97e9905fe81b509c44b589ce53c43f1f9f28634eb10e3375782c80a2f09e32c1"},"schema_version":"1.0","source":{"id":"2407.08733","kind":"arxiv","version":2}},"canonical_sha256":"b8da5cc4704d9d3ac094b3d37fc4be584a1602c976b0a623b8f04a8729c89249","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b8da5cc4704d9d3ac094b3d37fc4be584a1602c976b0a623b8f04a8729c89249","first_computed_at":"2026-07-05T09:17:45.829095Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:17:45.829095Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"pTSYvGye2QGiqh+Sl9Yzg2ny32BdRqZTbRbLvw8py76wPf2on7Mj3XmnScNgAVA7OkBlABZGfE68oAzLW4+KAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T09:17:45.829557Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.08733","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:64f396a7b4c21c75a5b11b672cca217ec2b1f6b9473ecab2cd3687ef0e738e3a","sha256:b84595f2d089d48843acf57031f71c30441c209658d94328bdcd37e4c01a4c4d"],"state_sha256":"0dbb4ca89416931b444e35f217e5541c08eb508c92ff7b37b7885a3d835d182a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"m85jGLAEW2rcypVViQ5JE8ixlaUvP1HNLLpXfpalkG6kXQ1/lqBK/RrG7zIXOHczHKzChFvzPQ3tWmpsyM9+AA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T20:52:03.521864Z","bundle_sha256":"406f7e6fca4c802559f49eb6691605b347b6fe81a901519bcc409a44a9e9b769"}}