{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MUWRI5DUNQZWJHIL5MWUGX7NHQ","short_pith_number":"pith:MUWRI5DU","schema_version":"1.0","canonical_sha256":"652d1474746c33649d0beb2d435fed3c243b29a2b2badded9cf6659dff4119c6","source":{"kind":"arxiv","id":"2501.12789","version":1},"attestation_state":"computed","paper":{"title":"Generating Diverse Q&A Benchmarks for RAG Evaluation with DataMorgana","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.IR"],"primary_cat":"cs.CL","authors_text":"David Carmel, Guy Horowitz, Liane Lewin-Eytan, Simone Filice, Yoelle Maarek, Zohar Karnin","submitted_at":"2025-01-22T10:47:08Z","abstract_excerpt":"Evaluating Retrieval-Augmented Generation (RAG) systems, especially in domain-specific contexts, requires benchmarks that address the distinctive requirements of the applicative scenario. Since real data can be hard to obtain, a common strategy is to use LLM-based methods to generate synthetic data. Existing solutions are general purpose: given a document, they generate a question to build a Q&A pair. However, although the generated questions can be individually good, they are typically not diverse enough to reasonably cover the different ways real end-users can interact with the RAG system. W"},"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":"2501.12789","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-22T10:47:08Z","cross_cats_sorted":["cs.IR"],"title_canon_sha256":"b5c4a303f9cfabd407ddacd9aed79e1f0859f5ab75713599667fe70dde034a5a","abstract_canon_sha256":"0dc0750ad592e52597c3118d277fc330a1e0dd4076bedfa9b2c579ea01ff9523"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:03:58.729833Z","signature_b64":"JNA6KDDLIHHLrB2ORnyTY3sfxO65vllc90nZ/iVHSKfNxvZPkvR416jke9FmPJGbwyhVSkGCn7OJQcdnW6ACAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"652d1474746c33649d0beb2d435fed3c243b29a2b2badded9cf6659dff4119c6","last_reissued_at":"2026-07-05T10:03:58.729338Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:03:58.729338Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generating Diverse Q&A Benchmarks for RAG Evaluation with DataMorgana","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.IR"],"primary_cat":"cs.CL","authors_text":"David Carmel, Guy Horowitz, Liane Lewin-Eytan, Simone Filice, Yoelle Maarek, Zohar Karnin","submitted_at":"2025-01-22T10:47:08Z","abstract_excerpt":"Evaluating Retrieval-Augmented Generation (RAG) systems, especially in domain-specific contexts, requires benchmarks that address the distinctive requirements of the applicative scenario. Since real data can be hard to obtain, a common strategy is to use LLM-based methods to generate synthetic data. Existing solutions are general purpose: given a document, they generate a question to build a Q&A pair. However, although the generated questions can be individually good, they are typically not diverse enough to reasonably cover the different ways real end-users can interact with the RAG system. W"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.12789","kind":"arxiv","version":1},"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/2501.12789/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":"2501.12789","created_at":"2026-07-05T10:03:58.729400+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.12789v1","created_at":"2026-07-05T10:03:58.729400+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.12789","created_at":"2026-07-05T10:03:58.729400+00:00"},{"alias_kind":"pith_short_12","alias_value":"MUWRI5DUNQZW","created_at":"2026-07-05T10:03:58.729400+00:00"},{"alias_kind":"pith_short_16","alias_value":"MUWRI5DUNQZWJHIL","created_at":"2026-07-05T10:03:58.729400+00:00"},{"alias_kind":"pith_short_8","alias_value":"MUWRI5DU","created_at":"2026-07-05T10:03:58.729400+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03650","citing_title":"CoEval: Ranking Language Models for Custom Tasks Without Labeled Data or Trustworthy Benchmarks","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12789","citing_title":"How Fine-Grained Should a RAG Benchmark Be? A Hierarchical Framework for Synthetic Question Generation","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2506.13743","citing_title":"LTRR: Learning To Rank Retrievers for LLMs","ref_index":18,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ","json":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ.json","graph_json":"https://pith.science/api/pith-number/MUWRI5DUNQZWJHIL5MWUGX7NHQ/graph.json","events_json":"https://pith.science/api/pith-number/MUWRI5DUNQZWJHIL5MWUGX7NHQ/events.json","paper":"https://pith.science/paper/MUWRI5DU"},"agent_actions":{"view_html":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ","download_json":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ.json","view_paper":"https://pith.science/paper/MUWRI5DU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.12789&json=true","fetch_graph":"https://pith.science/api/pith-number/MUWRI5DUNQZWJHIL5MWUGX7NHQ/graph.json","fetch_events":"https://pith.science/api/pith-number/MUWRI5DUNQZWJHIL5MWUGX7NHQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ/action/storage_attestation","attest_author":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ/action/author_attestation","sign_citation":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ/action/citation_signature","submit_replication":"https://pith.science/pith/MUWRI5DUNQZWJHIL5MWUGX7NHQ/action/replication_record"}},"created_at":"2026-07-05T10:03:58.729400+00:00","updated_at":"2026-07-05T10:03:58.729400+00:00"}