{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:L7XGLXUWLIBDHFWDNGZZGENKBR","short_pith_number":"pith:L7XGLXUW","schema_version":"1.0","canonical_sha256":"5fee65de965a023396c369b39311aa0c46a9d789eecf451fa0afe21e75f4c981","source":{"kind":"arxiv","id":"2409.13740","version":2},"attestation_state":"computed","paper":{"title":"Language agents achieve superhuman synthesis of scientific knowledge","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.IR","physics.soc-ph"],"primary_cat":"cs.CL","authors_text":"Andrew D. White, James D. Braza, Jon M. Laurent, Manvitha Ponnapati, Michaela Hinks, Michael D. Skarlinski, Michael J. Hammerling, Sam Cox, Samuel G. Rodriques","submitted_at":"2024-09-10T16:37:58Z","abstract_excerpt":"Language models are known to hallucinate incorrect information, and it is unclear if they are sufficiently accurate and reliable for use in scientific research. We developed a rigorous human-AI comparison methodology to evaluate language model agents on real-world literature search tasks covering information retrieval, summarization, and contradiction detection tasks. We show that PaperQA2, a frontier language model agent optimized for improved factuality, matches or exceeds subject matter expert performance on three realistic literature research tasks without any restrictions on humans (i.e.,"},"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":"2409.13740","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-10T16:37:58Z","cross_cats_sorted":["cs.AI","cs.IR","physics.soc-ph"],"title_canon_sha256":"e62a4a295bd2ed9269234d7e5cdabd21b174751348d584cb5bc1268b538987c7","abstract_canon_sha256":"83e73e6b3e9edf8760583d888fefa5b59c7b916bea8c3ed40ff74da4148a5bed"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:12:09.843443Z","signature_b64":"aAuPF+5O92EiSSk9jxOGxLyk+KR6ZavheMOM0Low308vh4wel/qIZ6fa54qmQYrr0VCIQ+rUbMvkpQ5Mqo/zCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5fee65de965a023396c369b39311aa0c46a9d789eecf451fa0afe21e75f4c981","last_reissued_at":"2026-07-05T09:12:09.842943Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:12:09.842943Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language agents achieve superhuman synthesis of scientific knowledge","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.IR","physics.soc-ph"],"primary_cat":"cs.CL","authors_text":"Andrew D. White, James D. Braza, Jon M. Laurent, Manvitha Ponnapati, Michaela Hinks, Michael D. Skarlinski, Michael J. Hammerling, Sam Cox, Samuel G. Rodriques","submitted_at":"2024-09-10T16:37:58Z","abstract_excerpt":"Language models are known to hallucinate incorrect information, and it is unclear if they are sufficiently accurate and reliable for use in scientific research. We developed a rigorous human-AI comparison methodology to evaluate language model agents on real-world literature search tasks covering information retrieval, summarization, and contradiction detection tasks. We show that PaperQA2, a frontier language model agent optimized for improved factuality, matches or exceeds subject matter expert performance on three realistic literature research tasks without any restrictions on humans (i.e.,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.13740","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/2409.13740/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":"2409.13740","created_at":"2026-07-05T09:12:09.843001+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.13740v2","created_at":"2026-07-05T09:12:09.843001+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.13740","created_at":"2026-07-05T09:12:09.843001+00:00"},{"alias_kind":"pith_short_12","alias_value":"L7XGLXUWLIBD","created_at":"2026-07-05T09:12:09.843001+00:00"},{"alias_kind":"pith_short_16","alias_value":"L7XGLXUWLIBDHFWD","created_at":"2026-07-05T09:12:09.843001+00:00"},{"alias_kind":"pith_short_8","alias_value":"L7XGLXUW","created_at":"2026-07-05T09:12:09.843001+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":32,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.17406","citing_title":"EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2606.24530","citing_title":"NatureBench: Can Coding Agents Match the Published SOTA of Nature-Family Papers?","ref_index":79,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27052","citing_title":"Human--LLM Collaboration Is Transforming Complexity Metrics in Scientific Texts","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26130","citing_title":"Thinking Like a Scientist? A Structural Study of LLM-Generated Research Methods","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01588","citing_title":"OrchestrXR: A Multi-Agent System for Idea-to-Prototype XR Study Authoring","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07454","citing_title":"PaperFlow: Profiling, Recommending, and Adapting Across Daily Paper Streams","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00597","citing_title":"Multi-Turn Agentic Scientific Literature Search via Workflow Induction","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01061","citing_title":"Agentic generation of verifiable rules for deterministic, self-expanding reaction classification","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31478","citing_title":"One Reflection Is Not Enough: Self-Correcting Autonomous Research via Multi-Hypothesis Failure Attribution","ref_index":112,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07022","citing_title":"Self-Driving Datasets: From 20 Million Papers to Nuanced Biomedical Knowledge at Scale","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29981","citing_title":"Hephaestus: Toward a Cybersecurity AI Scientist","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27610","citing_title":"Eliot: Interactively $\\underline{E}$xploring Fast-Changing Scientific $\\underline{Li}$terature Trends with $\\underline{O}$nline Da$\\underline{t}$a and Learning","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23204","citing_title":"AutoResearch AI: Towards AI-Powered Research Automation for Scientific Discovery","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2502.02871","citing_title":"Position: Multimodal Large Language Models Can Significantly Advance Scientific Reasoning","ref_index":171,"is_internal_anchor":false},{"citing_arxiv_id":"2502.13957","citing_title":"Supervising the search process produces reliable and generalizable information-seeking agents","ref_index":68,"is_internal_anchor":false},{"citing_arxiv_id":"2511.02824","citing_title":"Kosmos: An AI Scientist for Autonomous Discovery","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00505","citing_title":"LLM-Oriented Information Retrieval: A Denoising-First Perspective","ref_index":172,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07022","citing_title":"Self-Driving Datasets: From 20 Million Papers to Nuanced Biomedical Knowledge at Scale","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18490","citing_title":"Vector RAG vs LLM-Compiled Wiki: A Preregistered Comparison on a Small Multi-Domain Research","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18144","citing_title":"Evidence-Grounded Frontier Mapping and Agentic Hypothesis Generation in Nanomedicine","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18661","citing_title":"AI for Auto-Research: Roadmap & User Guide","ref_index":190,"is_internal_anchor":false},{"citing_arxiv_id":"2506.22598","citing_title":"RExBench: Can coding agents autonomously implement AI research extensions?","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2507.11810","citing_title":"Evolving Roles of LLMs in Scientific Innovation: Assistant, Collaborator, Scientist, and Evaluator","ref_index":159,"is_internal_anchor":false},{"citing_arxiv_id":"2509.26574","citing_title":"Probing the Critical Point (CritPt) of AI Reasoning: a Frontier Physics Research Benchmark","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2601.15170","citing_title":"Multi-Dimensional Knowledge Profiling with Large-Scale Literature Database and Hierarchical Retrieval","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR","json":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR.json","graph_json":"https://pith.science/api/pith-number/L7XGLXUWLIBDHFWDNGZZGENKBR/graph.json","events_json":"https://pith.science/api/pith-number/L7XGLXUWLIBDHFWDNGZZGENKBR/events.json","paper":"https://pith.science/paper/L7XGLXUW"},"agent_actions":{"view_html":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR","download_json":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR.json","view_paper":"https://pith.science/paper/L7XGLXUW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.13740&json=true","fetch_graph":"https://pith.science/api/pith-number/L7XGLXUWLIBDHFWDNGZZGENKBR/graph.json","fetch_events":"https://pith.science/api/pith-number/L7XGLXUWLIBDHFWDNGZZGENKBR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR/action/storage_attestation","attest_author":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR/action/author_attestation","sign_citation":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR/action/citation_signature","submit_replication":"https://pith.science/pith/L7XGLXUWLIBDHFWDNGZZGENKBR/action/replication_record"}},"created_at":"2026-07-05T09:12:09.843001+00:00","updated_at":"2026-07-05T09:12:09.843001+00:00"}