{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:AS7HDM7E4BQKWUALXJPBMZ52DC","short_pith_number":"pith:AS7HDM7E","schema_version":"1.0","canonical_sha256":"04be71b3e4e060ab500bba5e1667ba188b913c14755101517b15c22808ff534b","source":{"kind":"arxiv","id":"2508.15851","version":2},"attestation_state":"computed","paper":{"title":"DocHop-QA: Towards Multi-Hop Reasoning over Multimodal Document Collections","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jinwoo Kim, Jiwon Park, Rina Carines Cabal, Seohyun Pyeon, Soyeon Caren Han, Yihao Ding, Zhenyuan He","submitted_at":"2025-08-20T08:17:45Z","abstract_excerpt":"Despite rapid progress in large language models (LLMs), current QA benchmarks still overlook the core challenge of real-world scientific information seeking: synthesizing multimodal evidence scattered across multiple documents and structural formats. Existing QA benchmarks remain narrow in scope, relying on unimodal text and short-span reasoning that fail to capture the complexity of real information seeking. We introduce DocHop-QA, a benchmark of 11,379 instances for evaluating multimodal, multi-document, multi-hop scientific QA. Built from publicly available PubMed articles, DocHop-QA incorp"},"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":"2508.15851","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-20T08:17:45Z","cross_cats_sorted":[],"title_canon_sha256":"baddfb8d0d1b77131aa72d2ff843920b70164daeb255013f17e0661e1b5ab90c","abstract_canon_sha256":"22aee27774e31cd45074463c56de7a5dc4745195c96561575256a760bea526c9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:15:15.511860Z","signature_b64":"Aa6/Qt1ywZKq3swcuNMz8kpwcxUSDwKgzXPqMeurtkZaQWaoK7qPdG0nj+zgjNe6PJhMATHx7capTal6E9TUBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"04be71b3e4e060ab500bba5e1667ba188b913c14755101517b15c22808ff534b","last_reissued_at":"2026-06-05T01:15:15.511107Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:15:15.511107Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DocHop-QA: Towards Multi-Hop Reasoning over Multimodal Document Collections","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jinwoo Kim, Jiwon Park, Rina Carines Cabal, Seohyun Pyeon, Soyeon Caren Han, Yihao Ding, Zhenyuan He","submitted_at":"2025-08-20T08:17:45Z","abstract_excerpt":"Despite rapid progress in large language models (LLMs), current QA benchmarks still overlook the core challenge of real-world scientific information seeking: synthesizing multimodal evidence scattered across multiple documents and structural formats. Existing QA benchmarks remain narrow in scope, relying on unimodal text and short-span reasoning that fail to capture the complexity of real information seeking. We introduce DocHop-QA, a benchmark of 11,379 instances for evaluating multimodal, multi-document, multi-hop scientific QA. Built from publicly available PubMed articles, DocHop-QA incorp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.15851","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/2508.15851/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":"2508.15851","created_at":"2026-06-05T01:15:15.511198+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.15851v2","created_at":"2026-06-05T01:15:15.511198+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.15851","created_at":"2026-06-05T01:15:15.511198+00:00"},{"alias_kind":"pith_short_12","alias_value":"AS7HDM7E4BQK","created_at":"2026-06-05T01:15:15.511198+00:00"},{"alias_kind":"pith_short_16","alias_value":"AS7HDM7E4BQKWUAL","created_at":"2026-06-05T01:15:15.511198+00:00"},{"alias_kind":"pith_short_8","alias_value":"AS7HDM7E","created_at":"2026-06-05T01:15:15.511198+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC","json":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC.json","graph_json":"https://pith.science/api/pith-number/AS7HDM7E4BQKWUALXJPBMZ52DC/graph.json","events_json":"https://pith.science/api/pith-number/AS7HDM7E4BQKWUALXJPBMZ52DC/events.json","paper":"https://pith.science/paper/AS7HDM7E"},"agent_actions":{"view_html":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC","download_json":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC.json","view_paper":"https://pith.science/paper/AS7HDM7E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.15851&json=true","fetch_graph":"https://pith.science/api/pith-number/AS7HDM7E4BQKWUALXJPBMZ52DC/graph.json","fetch_events":"https://pith.science/api/pith-number/AS7HDM7E4BQKWUALXJPBMZ52DC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC/action/storage_attestation","attest_author":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC/action/author_attestation","sign_citation":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC/action/citation_signature","submit_replication":"https://pith.science/pith/AS7HDM7E4BQKWUALXJPBMZ52DC/action/replication_record"}},"created_at":"2026-06-05T01:15:15.511198+00:00","updated_at":"2026-06-05T01:15:15.511198+00:00"}