{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:CLESLWG5UDG3G4I66G6NGHWQO7","short_pith_number":"pith:CLESLWG5","canonical_record":{"source":{"id":"2305.12524","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-21T17:51:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c9b26605822aab58426a431df593154eb8db8cb69fec500c34a4f698936b92a4","abstract_canon_sha256":"7109df58195b71a72333019bef8cce2485d61bab0e61f8c6fa5cc64dd8fbb6e6"},"schema_version":"1.0"},"canonical_sha256":"12c925d8dda0cdb3711ef1bcd31ed077efd118a45f01956d83e6e59c586e3038","source":{"kind":"arxiv","id":"2305.12524","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.12524","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"arxiv_version","alias_value":"2305.12524v3","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.12524","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"pith_short_12","alias_value":"CLESLWG5UDG3","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"pith_short_16","alias_value":"CLESLWG5UDG3G4I6","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"pith_short_8","alias_value":"CLESLWG5","created_at":"2026-07-05T07:20:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:CLESLWG5UDG3G4I66G6NGHWQO7","target":"record","payload":{"canonical_record":{"source":{"id":"2305.12524","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-21T17:51:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c9b26605822aab58426a431df593154eb8db8cb69fec500c34a4f698936b92a4","abstract_canon_sha256":"7109df58195b71a72333019bef8cce2485d61bab0e61f8c6fa5cc64dd8fbb6e6"},"schema_version":"1.0"},"canonical_sha256":"12c925d8dda0cdb3711ef1bcd31ed077efd118a45f01956d83e6e59c586e3038","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:20:50.950659Z","signature_b64":"9yJ9839/GasR/tDRcL5Bu8795mQjlCmsMOsWjLuJs4XKc1tZAPu+fuHSo7yAeSABPhH/MVT64xfpyZVtdHMPAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"12c925d8dda0cdb3711ef1bcd31ed077efd118a45f01956d83e6e59c586e3038","last_reissued_at":"2026-07-05T07:20:50.950164Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:20:50.950164Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2305.12524","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:20:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Tlb+mPpom2fk/yCvl008yGzzZ56kb9CIMPa1jVyXWq/6XRuQlZJChPZyE3hS7HJdfGgAV42hp2jPPFpURrbsCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T22:27:25.061926Z"},"content_sha256":"fd8bc5ba07a411886ab52ceadd125c05b16ee5039a66e89f46fad5d7c9d77d69","schema_version":"1.0","event_id":"sha256:fd8bc5ba07a411886ab52ceadd125c05b16ee5039a66e89f46fad5d7c9d77d69"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:CLESLWG5UDG3G4I66G6NGHWQO7","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"TheoremQA: A Theorem-driven Question Answering dataset","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jianyu Xu, Max Ku, Ming Yin, Pan Lu, Tony Xia, Wenhu Chen, Xinyi Wang, Xueguang Ma, Yixin Wan","submitted_at":"2023-05-21T17:51:35Z","abstract_excerpt":"The recent LLMs like GPT-4 and PaLM-2 have made tremendous progress in solving fundamental math problems like GSM8K by achieving over 90% accuracy. However, their capabilities to solve more challenging math problems which require domain-specific knowledge (i.e. theorem) have yet to be investigated. In this paper, we introduce TheoremQA, the first theorem-driven question-answering dataset designed to evaluate AI models' capabilities to apply theorems to solve challenging science problems. TheoremQA is curated by domain experts containing 800 high-quality questions covering 350 theorems (e.g. Ta"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.12524","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/2305.12524/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":"2026-07-05T07:20:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Ot7nDcfsV8kJaOcXIxnc1D4Kj2JxTUcx3tSrjimUDJat5Zehep2Ra8DB/pDCF4dJOv9bvD0EvcLLuM2LEDJeDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T22:27:25.062790Z"},"content_sha256":"1050a43d4317ff19248bc39511b1c42c2b9e14f59129467f5094966670ce051d","schema_version":"1.0","event_id":"sha256:1050a43d4317ff19248bc39511b1c42c2b9e14f59129467f5094966670ce051d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CLESLWG5UDG3G4I66G6NGHWQO7/bundle.json","state_url":"https://pith.science/pith/CLESLWG5UDG3G4I66G6NGHWQO7/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CLESLWG5UDG3G4I66G6NGHWQO7/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":"2026-08-08T22:27:25Z","links":{"resolver":"https://pith.science/pith/CLESLWG5UDG3G4I66G6NGHWQO7","bundle":"https://pith.science/pith/CLESLWG5UDG3G4I66G6NGHWQO7/bundle.json","state":"https://pith.science/pith/CLESLWG5UDG3G4I66G6NGHWQO7/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CLESLWG5UDG3G4I66G6NGHWQO7/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:CLESLWG5UDG3G4I66G6NGHWQO7","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":"7109df58195b71a72333019bef8cce2485d61bab0e61f8c6fa5cc64dd8fbb6e6","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-21T17:51:35Z","title_canon_sha256":"c9b26605822aab58426a431df593154eb8db8cb69fec500c34a4f698936b92a4"},"schema_version":"1.0","source":{"id":"2305.12524","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.12524","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"arxiv_version","alias_value":"2305.12524v3","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.12524","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"pith_short_12","alias_value":"CLESLWG5UDG3","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"pith_short_16","alias_value":"CLESLWG5UDG3G4I6","created_at":"2026-07-05T07:20:50Z"},{"alias_kind":"pith_short_8","alias_value":"CLESLWG5","created_at":"2026-07-05T07:20:50Z"}],"graph_snapshots":[{"event_id":"sha256:1050a43d4317ff19248bc39511b1c42c2b9e14f59129467f5094966670ce051d","target":"graph","created_at":"2026-07-05T07:20:50Z","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/2305.12524/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The recent LLMs like GPT-4 and PaLM-2 have made tremendous progress in solving fundamental math problems like GSM8K by achieving over 90% accuracy. However, their capabilities to solve more challenging math problems which require domain-specific knowledge (i.e. theorem) have yet to be investigated. In this paper, we introduce TheoremQA, the first theorem-driven question-answering dataset designed to evaluate AI models' capabilities to apply theorems to solve challenging science problems. TheoremQA is curated by domain experts containing 800 high-quality questions covering 350 theorems (e.g. Ta","authors_text":"Jianyu Xu, Max Ku, Ming Yin, Pan Lu, Tony Xia, Wenhu Chen, Xinyi Wang, Xueguang Ma, Yixin Wan","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-21T17:51:35Z","title":"TheoremQA: A Theorem-driven Question Answering dataset"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.12524","kind":"arxiv","version":3},"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:fd8bc5ba07a411886ab52ceadd125c05b16ee5039a66e89f46fad5d7c9d77d69","target":"record","created_at":"2026-07-05T07:20:50Z","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":"7109df58195b71a72333019bef8cce2485d61bab0e61f8c6fa5cc64dd8fbb6e6","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-21T17:51:35Z","title_canon_sha256":"c9b26605822aab58426a431df593154eb8db8cb69fec500c34a4f698936b92a4"},"schema_version":"1.0","source":{"id":"2305.12524","kind":"arxiv","version":3}},"canonical_sha256":"12c925d8dda0cdb3711ef1bcd31ed077efd118a45f01956d83e6e59c586e3038","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"12c925d8dda0cdb3711ef1bcd31ed077efd118a45f01956d83e6e59c586e3038","first_computed_at":"2026-07-05T07:20:50.950164Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:20:50.950164Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"9yJ9839/GasR/tDRcL5Bu8795mQjlCmsMOsWjLuJs4XKc1tZAPu+fuHSo7yAeSABPhH/MVT64xfpyZVtdHMPAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T07:20:50.950659Z","signed_message":"canonical_sha256_bytes"},"source_id":"2305.12524","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fd8bc5ba07a411886ab52ceadd125c05b16ee5039a66e89f46fad5d7c9d77d69","sha256:1050a43d4317ff19248bc39511b1c42c2b9e14f59129467f5094966670ce051d"],"state_sha256":"7465b83facd71167d76de8e0424cde5e014801b72d9695060fdb5c8cddaa659b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6wm/tC+Y8rKytdgkUZFbInXkmBEvj8UuIZVteNpHZB3o6vMfPEDeIvt4Sdv54gy+/Uff2zqw1TPfjOdBOjL7Bw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T22:27:25.069476Z","bundle_sha256":"ee062dbfa3cc9857e81dae227c036df0d57c8c20f78b3c278aa81a186a79725e"}}