{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:L2DGR7N2MRERBFP25PY3R3MJ62","short_pith_number":"pith:L2DGR7N2","schema_version":"1.0","canonical_sha256":"5e8668fdba64491095faebf1b8ed89f685af12679384d0cadd89759ae15c145c","source":{"kind":"arxiv","id":"2410.13961","version":2},"attestation_state":"computed","paper":{"title":"From Single to Multi: How LLMs Hallucinate in Multi-Document Summarization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Catarina G. Belem, Estevam Hruschka, Hayate Iso, Nikita Bhutani, Pouya Pezeshkpour, Seiji Maekawa","submitted_at":"2024-10-17T18:38:53Z","abstract_excerpt":"Although many studies have investigated and reduced hallucinations in large language models (LLMs) for single-document tasks, research on hallucination in multi-document summarization (MDS) tasks remains largely unexplored. Specifically, it is unclear how the challenges arising from handling multiple documents (e.g., repetition and diversity of information) affect models outputs. In this work, we investigate how hallucinations manifest in LLMs when summarizing topic-specific information from multiple documents. Since no benchmarks exist for investigating hallucinations in MDS, we use existing "},"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":"2410.13961","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-17T18:38:53Z","cross_cats_sorted":[],"title_canon_sha256":"74b5e46222925487cdfa13742dcb7b381f79a6a69edc5254aee5366680efe9b2","abstract_canon_sha256":"095b729c337b9b569468a02137439f0c26ce38a0ddd43d3be091f1f0c5746025"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:54:13.427451Z","signature_b64":"2Zuag7aphnePGaeqObOrBJXwf6LYJ3msMoaMYjTBBOAPHXwNxZaMNOmofYLefbsDJ9nGmbon3w3d4AOBirZFDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5e8668fdba64491095faebf1b8ed89f685af12679384d0cadd89759ae15c145c","last_reissued_at":"2026-07-05T10:54:13.426915Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:54:13.426915Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Single to Multi: How LLMs Hallucinate in Multi-Document Summarization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Catarina G. Belem, Estevam Hruschka, Hayate Iso, Nikita Bhutani, Pouya Pezeshkpour, Seiji Maekawa","submitted_at":"2024-10-17T18:38:53Z","abstract_excerpt":"Although many studies have investigated and reduced hallucinations in large language models (LLMs) for single-document tasks, research on hallucination in multi-document summarization (MDS) tasks remains largely unexplored. Specifically, it is unclear how the challenges arising from handling multiple documents (e.g., repetition and diversity of information) affect models outputs. In this work, we investigate how hallucinations manifest in LLMs when summarizing topic-specific information from multiple documents. Since no benchmarks exist for investigating hallucinations in MDS, we use existing "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.13961","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/2410.13961/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":"2410.13961","created_at":"2026-07-05T10:54:13.426977+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.13961v2","created_at":"2026-07-05T10:54:13.426977+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.13961","created_at":"2026-07-05T10:54:13.426977+00:00"},{"alias_kind":"pith_short_12","alias_value":"L2DGR7N2MRER","created_at":"2026-07-05T10:54:13.426977+00:00"},{"alias_kind":"pith_short_16","alias_value":"L2DGR7N2MRERBFP2","created_at":"2026-07-05T10:54:13.426977+00:00"},{"alias_kind":"pith_short_8","alias_value":"L2DGR7N2","created_at":"2026-07-05T10:54:13.426977+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2510.00361","citing_title":"Attribution Gradients: Incrementally Unfolding Citations for Critical Examination of Attributed AI Answers","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2505.06120","citing_title":"LLMs Get Lost In Multi-Turn Conversation","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62","json":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62.json","graph_json":"https://pith.science/api/pith-number/L2DGR7N2MRERBFP25PY3R3MJ62/graph.json","events_json":"https://pith.science/api/pith-number/L2DGR7N2MRERBFP25PY3R3MJ62/events.json","paper":"https://pith.science/paper/L2DGR7N2"},"agent_actions":{"view_html":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62","download_json":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62.json","view_paper":"https://pith.science/paper/L2DGR7N2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.13961&json=true","fetch_graph":"https://pith.science/api/pith-number/L2DGR7N2MRERBFP25PY3R3MJ62/graph.json","fetch_events":"https://pith.science/api/pith-number/L2DGR7N2MRERBFP25PY3R3MJ62/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62/action/storage_attestation","attest_author":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62/action/author_attestation","sign_citation":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62/action/citation_signature","submit_replication":"https://pith.science/pith/L2DGR7N2MRERBFP25PY3R3MJ62/action/replication_record"}},"created_at":"2026-07-05T10:54:13.426977+00:00","updated_at":"2026-07-05T10:54:13.426977+00:00"}