{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:LHNRSOHY45H22LWP5DKGFWK4NX","short_pith_number":"pith:LHNRSOHY","canonical_record":{"source":{"id":"2503.06218","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-08T13:40:10Z","cross_cats_sorted":[],"title_canon_sha256":"58eb68a791d04478f61e55c1fa32db1fc8a5123e52c8388407dd2c38fe1971ea","abstract_canon_sha256":"a98a35b2f08503c63c4b99c6cf906473bd815a4e223085e48e8ef0580ee78f35"},"schema_version":"1.0"},"canonical_sha256":"59db1938f8e74fad2ecfe8d462d95c6de0c424aef0d3f63e9cd62805dc6a22d8","source":{"kind":"arxiv","id":"2503.06218","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.06218","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"arxiv_version","alias_value":"2503.06218v2","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06218","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"pith_short_12","alias_value":"LHNRSOHY45H2","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"pith_short_16","alias_value":"LHNRSOHY45H22LWP","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"pith_short_8","alias_value":"LHNRSOHY","created_at":"2026-07-05T11:04:46Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:LHNRSOHY45H22LWP5DKGFWK4NX","target":"record","payload":{"canonical_record":{"source":{"id":"2503.06218","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-08T13:40:10Z","cross_cats_sorted":[],"title_canon_sha256":"58eb68a791d04478f61e55c1fa32db1fc8a5123e52c8388407dd2c38fe1971ea","abstract_canon_sha256":"a98a35b2f08503c63c4b99c6cf906473bd815a4e223085e48e8ef0580ee78f35"},"schema_version":"1.0"},"canonical_sha256":"59db1938f8e74fad2ecfe8d462d95c6de0c424aef0d3f63e9cd62805dc6a22d8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:04:46.973939Z","signature_b64":"024DomfLgeuIeQyR5rRMDAkpQ7Z62CSjPrKPOz7lYjwqnoB0dex/TEMKCpQuiDmbX6Ae5B84/u2fy7DpDWfrAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"59db1938f8e74fad2ecfe8d462d95c6de0c424aef0d3f63e9cd62805dc6a22d8","last_reissued_at":"2026-07-05T11:04:46.973394Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:04:46.973394Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2503.06218","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":"2026-07-05T11:04:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7KRHjyj1Ju9w5VyEuSBcHHCvUXfFW0t0Jhb299kvFUpkIWgrdN3uPIPvBjaPKCNPP/V0716FBkuORFMT3DIAAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T15:40:38.303453Z"},"content_sha256":"b9ee629abbccf8581de21a1179a568d7b89eaa871aece6b10b15806b05a0a173","schema_version":"1.0","event_id":"sha256:b9ee629abbccf8581de21a1179a568d7b89eaa871aece6b10b15806b05a0a173"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:LHNRSOHY45H22LWP5DKGFWK4NX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SCoRE: Benchmarking Long-Chain Reasoning in Commonsense Scenarios","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jingyuan Ma, Jinkun Ding, Liming Xiao, Nan Hu, Qun Liu, Rui Li, Sirui Deng, Weidong Zhan, Weilin Luo, Wenhan Ma, Yixin Yang, Yue Wang, Yuhang Qin, Zheng Li, Zhifang Sui","submitted_at":"2025-03-08T13:40:10Z","abstract_excerpt":"Currently, long-chain reasoning remains a key challenge for large language models (LLMs) because natural texts lack sufficient explicit reasoning data. However, existing benchmarks suffer from limitations such as narrow coverage, short reasoning paths, or high construction costs. We introduce SCoRE (Scenario-based Commonsense Reasoning Evaluation), a benchmark that synthesizes multi-hop questions from scenario schemas of entities, relations, and logical rules to assess long-chain commonsense reasoning. SCoRE contains 100k bilingual (Chinese-English) multiple-choice questions whose reasoning ch"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06218","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/2503.06218/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":"2026-07-05T11:04:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7zCU2kKo3Kpp6uXORXNWRpwFaJ6It77U/CLkxjteTq2D8lOA1SnqA9bcgSNuZw/oqFGtsKTOXtoxPzs9MHpBBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T15:40:38.303967Z"},"content_sha256":"21b6a0d1f0c641030c2ff17b354d431945c70f991c75f8b39b99432e55487b45","schema_version":"1.0","event_id":"sha256:21b6a0d1f0c641030c2ff17b354d431945c70f991c75f8b39b99432e55487b45"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LHNRSOHY45H22LWP5DKGFWK4NX/bundle.json","state_url":"https://pith.science/pith/LHNRSOHY45H22LWP5DKGFWK4NX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LHNRSOHY45H22LWP5DKGFWK4NX/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":"2026-08-05T15:40:38Z","links":{"resolver":"https://pith.science/pith/LHNRSOHY45H22LWP5DKGFWK4NX","bundle":"https://pith.science/pith/LHNRSOHY45H22LWP5DKGFWK4NX/bundle.json","state":"https://pith.science/pith/LHNRSOHY45H22LWP5DKGFWK4NX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LHNRSOHY45H22LWP5DKGFWK4NX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:LHNRSOHY45H22LWP5DKGFWK4NX","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":"a98a35b2f08503c63c4b99c6cf906473bd815a4e223085e48e8ef0580ee78f35","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-08T13:40:10Z","title_canon_sha256":"58eb68a791d04478f61e55c1fa32db1fc8a5123e52c8388407dd2c38fe1971ea"},"schema_version":"1.0","source":{"id":"2503.06218","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.06218","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"arxiv_version","alias_value":"2503.06218v2","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06218","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"pith_short_12","alias_value":"LHNRSOHY45H2","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"pith_short_16","alias_value":"LHNRSOHY45H22LWP","created_at":"2026-07-05T11:04:46Z"},{"alias_kind":"pith_short_8","alias_value":"LHNRSOHY","created_at":"2026-07-05T11:04:46Z"}],"graph_snapshots":[{"event_id":"sha256:21b6a0d1f0c641030c2ff17b354d431945c70f991c75f8b39b99432e55487b45","target":"graph","created_at":"2026-07-05T11:04:46Z","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/2503.06218/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Currently, long-chain reasoning remains a key challenge for large language models (LLMs) because natural texts lack sufficient explicit reasoning data. However, existing benchmarks suffer from limitations such as narrow coverage, short reasoning paths, or high construction costs. We introduce SCoRE (Scenario-based Commonsense Reasoning Evaluation), a benchmark that synthesizes multi-hop questions from scenario schemas of entities, relations, and logical rules to assess long-chain commonsense reasoning. SCoRE contains 100k bilingual (Chinese-English) multiple-choice questions whose reasoning ch","authors_text":"Jingyuan Ma, Jinkun Ding, Liming Xiao, Nan Hu, Qun Liu, Rui Li, Sirui Deng, Weidong Zhan, Weilin Luo, Wenhan Ma, Yixin Yang, Yue Wang, Yuhang Qin, Zheng Li, Zhifang Sui","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-08T13:40:10Z","title":"SCoRE: Benchmarking Long-Chain Reasoning in Commonsense Scenarios"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06218","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:b9ee629abbccf8581de21a1179a568d7b89eaa871aece6b10b15806b05a0a173","target":"record","created_at":"2026-07-05T11:04:46Z","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":"a98a35b2f08503c63c4b99c6cf906473bd815a4e223085e48e8ef0580ee78f35","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-08T13:40:10Z","title_canon_sha256":"58eb68a791d04478f61e55c1fa32db1fc8a5123e52c8388407dd2c38fe1971ea"},"schema_version":"1.0","source":{"id":"2503.06218","kind":"arxiv","version":2}},"canonical_sha256":"59db1938f8e74fad2ecfe8d462d95c6de0c424aef0d3f63e9cd62805dc6a22d8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"59db1938f8e74fad2ecfe8d462d95c6de0c424aef0d3f63e9cd62805dc6a22d8","first_computed_at":"2026-07-05T11:04:46.973394Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:04:46.973394Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"024DomfLgeuIeQyR5rRMDAkpQ7Z62CSjPrKPOz7lYjwqnoB0dex/TEMKCpQuiDmbX6Ae5B84/u2fy7DpDWfrAw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:04:46.973939Z","signed_message":"canonical_sha256_bytes"},"source_id":"2503.06218","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b9ee629abbccf8581de21a1179a568d7b89eaa871aece6b10b15806b05a0a173","sha256:21b6a0d1f0c641030c2ff17b354d431945c70f991c75f8b39b99432e55487b45"],"state_sha256":"a2495d170f13fc299c70a13727375158038fc54cd22be6daaf958d1a68a43a45"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Zvqo7W+7m3CfCufZdaxadiL+HhbWd7h5cQ+h/oLUETIjXEMmSqsODRzBt32yTpPNa73szhbem9vlTnkRPb1HBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T15:40:38.307218Z","bundle_sha256":"3a9ef7052135ae53b297270f9958880111a7e2803645d125f2ecde07c99e4dec"}}