{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:XLF2IVXDUXH6FTO6YFS3C4JXF3","short_pith_number":"pith:XLF2IVXD","canonical_record":{"source":{"id":"2410.22394","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T17:58:29Z","cross_cats_sorted":[],"title_canon_sha256":"77edb57d34e7509015c9ff8a194ba5a2c2f4f1d6f6ffeb48e1fbabd298be16cf","abstract_canon_sha256":"62026c41efee5d4368da95114be4de1f2e043a6761cc9950b0ebf976deff375d"},"schema_version":"1.0"},"canonical_sha256":"bacba456e3a5cfe2cddec165b171372edf1fc9f386d2214448a89fdeafceaf09","source":{"kind":"arxiv","id":"2410.22394","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.22394","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"arxiv_version","alias_value":"2410.22394v4","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.22394","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"pith_short_12","alias_value":"XLF2IVXDUXH6","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"pith_short_16","alias_value":"XLF2IVXDUXH6FTO6","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"pith_short_8","alias_value":"XLF2IVXD","created_at":"2026-07-05T11:09:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:XLF2IVXDUXH6FTO6YFS3C4JXF3","target":"record","payload":{"canonical_record":{"source":{"id":"2410.22394","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T17:58:29Z","cross_cats_sorted":[],"title_canon_sha256":"77edb57d34e7509015c9ff8a194ba5a2c2f4f1d6f6ffeb48e1fbabd298be16cf","abstract_canon_sha256":"62026c41efee5d4368da95114be4de1f2e043a6761cc9950b0ebf976deff375d"},"schema_version":"1.0"},"canonical_sha256":"bacba456e3a5cfe2cddec165b171372edf1fc9f386d2214448a89fdeafceaf09","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:09:02.470151Z","signature_b64":"WgpzviKLJFXNIKqrN/UiFfGBzQLkdwVQ2AMyw+Fj/wg+a+fMx2NOd+eJobpv1EvCSmkFaaywWS+Qlg7yc89wDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bacba456e3a5cfe2cddec165b171372edf1fc9f386d2214448a89fdeafceaf09","last_reissued_at":"2026-07-05T11:09:02.469696Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:09:02.469696Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.22394","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:09:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"b0UnIXtwZDnzs7klk3kixOIXDbwnOou9PiNnSgef92eNPBGnjbpAY8a1aIQQ4JW790wrMLZdPn198f0q+BMXCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T16:44:28.459456Z"},"content_sha256":"a6ae2b0753f8f7efec358c0f4e9f830aad31bd7a13be4dad8c79e3e12303aac5","schema_version":"1.0","event_id":"sha256:a6ae2b0753f8f7efec358c0f4e9f830aad31bd7a13be4dad8c79e3e12303aac5"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:XLF2IVXDUXH6FTO6YFS3C4JXF3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AAAR-1.0: Assessing AI's Potential to Assist Research","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Congying Xia, Hanzi Xu, Hongchao Fang, Jiangshu Du, Jian Xie, Jihyun Janice Ahn, Kai Zhang, Lifu Huang, Renze Lou, Ryo Kamoi, Sijia Wang, Wenchao Ma, Wenpeng Yin, Xiaoxin Lu, Xi Li, Yusen Zhang, Yuxuan Sun, Zhuoyang Zou","submitted_at":"2024-10-29T17:58:29Z","abstract_excerpt":"Numerous studies have assessed the proficiency of AI systems, particularly large language models (LLMs), in facilitating everyday tasks such as email writing, question answering, and creative content generation. However, researchers face unique challenges and opportunities in leveraging LLMs for their own work, such as brainstorming research ideas, designing experiments, and writing or reviewing papers. In this study, we introduce AAAR-1.0, a benchmark dataset designed to evaluate LLM performance in three fundamental, expertise-intensive research tasks: (i) EquationInference, assessing the cor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.22394","kind":"arxiv","version":4},"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/2410.22394/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":"2026-07-05T11:09:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZU19kPdYXNgLMuyGB8Chi6JxP0iR6zc3wZmJzyvsejClQXGfxH0bX1J0AjryiDBfkkkXGS7pBGmliVqb4kiQCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T16:44:28.459996Z"},"content_sha256":"08c1a68765be595675f696814148324e5fb9db382376b7e074377344f20172e1","schema_version":"1.0","event_id":"sha256:08c1a68765be595675f696814148324e5fb9db382376b7e074377344f20172e1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/XLF2IVXDUXH6FTO6YFS3C4JXF3/bundle.json","state_url":"https://pith.science/pith/XLF2IVXDUXH6FTO6YFS3C4JXF3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/XLF2IVXDUXH6FTO6YFS3C4JXF3/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":"2026-08-10T16:44:28Z","links":{"resolver":"https://pith.science/pith/XLF2IVXDUXH6FTO6YFS3C4JXF3","bundle":"https://pith.science/pith/XLF2IVXDUXH6FTO6YFS3C4JXF3/bundle.json","state":"https://pith.science/pith/XLF2IVXDUXH6FTO6YFS3C4JXF3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/XLF2IVXDUXH6FTO6YFS3C4JXF3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:XLF2IVXDUXH6FTO6YFS3C4JXF3","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":"62026c41efee5d4368da95114be4de1f2e043a6761cc9950b0ebf976deff375d","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T17:58:29Z","title_canon_sha256":"77edb57d34e7509015c9ff8a194ba5a2c2f4f1d6f6ffeb48e1fbabd298be16cf"},"schema_version":"1.0","source":{"id":"2410.22394","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.22394","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"arxiv_version","alias_value":"2410.22394v4","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.22394","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"pith_short_12","alias_value":"XLF2IVXDUXH6","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"pith_short_16","alias_value":"XLF2IVXDUXH6FTO6","created_at":"2026-07-05T11:09:02Z"},{"alias_kind":"pith_short_8","alias_value":"XLF2IVXD","created_at":"2026-07-05T11:09:02Z"}],"graph_snapshots":[{"event_id":"sha256:08c1a68765be595675f696814148324e5fb9db382376b7e074377344f20172e1","target":"graph","created_at":"2026-07-05T11:09:02Z","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/2410.22394/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Numerous studies have assessed the proficiency of AI systems, particularly large language models (LLMs), in facilitating everyday tasks such as email writing, question answering, and creative content generation. However, researchers face unique challenges and opportunities in leveraging LLMs for their own work, such as brainstorming research ideas, designing experiments, and writing or reviewing papers. In this study, we introduce AAAR-1.0, a benchmark dataset designed to evaluate LLM performance in three fundamental, expertise-intensive research tasks: (i) EquationInference, assessing the cor","authors_text":"Congying Xia, Hanzi Xu, Hongchao Fang, Jiangshu Du, Jian Xie, Jihyun Janice Ahn, Kai Zhang, Lifu Huang, Renze Lou, Ryo Kamoi, Sijia Wang, Wenchao Ma, Wenpeng Yin, Xiaoxin Lu, Xi Li, Yusen Zhang, Yuxuan Sun, Zhuoyang Zou","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T17:58:29Z","title":"AAAR-1.0: Assessing AI's Potential to Assist Research"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.22394","kind":"arxiv","version":4},"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:a6ae2b0753f8f7efec358c0f4e9f830aad31bd7a13be4dad8c79e3e12303aac5","target":"record","created_at":"2026-07-05T11:09:02Z","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":"62026c41efee5d4368da95114be4de1f2e043a6761cc9950b0ebf976deff375d","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T17:58:29Z","title_canon_sha256":"77edb57d34e7509015c9ff8a194ba5a2c2f4f1d6f6ffeb48e1fbabd298be16cf"},"schema_version":"1.0","source":{"id":"2410.22394","kind":"arxiv","version":4}},"canonical_sha256":"bacba456e3a5cfe2cddec165b171372edf1fc9f386d2214448a89fdeafceaf09","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"bacba456e3a5cfe2cddec165b171372edf1fc9f386d2214448a89fdeafceaf09","first_computed_at":"2026-07-05T11:09:02.469696Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:09:02.469696Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"WgpzviKLJFXNIKqrN/UiFfGBzQLkdwVQ2AMyw+Fj/wg+a+fMx2NOd+eJobpv1EvCSmkFaaywWS+Qlg7yc89wDw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:09:02.470151Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.22394","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a6ae2b0753f8f7efec358c0f4e9f830aad31bd7a13be4dad8c79e3e12303aac5","sha256:08c1a68765be595675f696814148324e5fb9db382376b7e074377344f20172e1"],"state_sha256":"614649fc8dec39914fe5c67eb2c97b47d9472df7bdf423565ea91ce7b444c207"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7X5WNrHuewoaJpYuRg1QtN06u4Z5V7L1G31K5EZ/ejhvuAQC1g2NYC8S4PVF+qCPyEv9eBZ4TUO9kfNgeF+vBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T16:44:28.463402Z","bundle_sha256":"0f4e3d79a16c55204a5e6feaf3b023d3efcaa2a1ec13e2f5adddebda6554185c"}}