{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KQNPUH37OGA5FVTFIJM7564V6G","short_pith_number":"pith:KQNPUH37","schema_version":"1.0","canonical_sha256":"541afa1f7f7181d2d6654259fefb95f1891a2efb747ed10761cdfe1cde12d81a","source":{"kind":"arxiv","id":"2401.10186","version":3},"attestation_state":"computed","paper":{"title":"Beyond Traditional Benchmarks: Analyzing Behaviors of Open LLMs on Data-to-Text Generation","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ond\\v{r}ej Du\\v{s}ek, Zden\\v{e}k Kasner","submitted_at":"2024-01-18T18:15:46Z","abstract_excerpt":"We analyze the behaviors of open large language models (LLMs) on the task of data-to-text (D2T) generation, i.e., generating coherent and relevant text from structured data. To avoid the issue of LLM training data contamination with standard benchmarks, we design Quintd - a tool for collecting novel structured data records from public APIs. We find that open LLMs (Llama 2, Mistral, and Zephyr) can generate fluent and coherent texts in zero-shot settings from data in common formats collected with Quintd. However, we show that the semantic accuracy of the outputs is a major issue: both according"},"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":"2401.10186","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-01-18T18:15:46Z","cross_cats_sorted":[],"title_canon_sha256":"233e32f53730385f746a9fde4a6b86adbe3d3435ed49a42c3572dad5f330b960","abstract_canon_sha256":"e2bcc89cc753918ba1ff266edae656ab3e429ae0a7c86a47ff9c7aa05c1cb8ff"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:05.091283Z","signature_b64":"0XY+AF3qdMawjVW2xJp1gWLriCHcaJd7K7HM3G8B5/j685ENw/x1YbpLSwXIBXl2K/+5Ol5y8LryPLXEHQWBAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"541afa1f7f7181d2d6654259fefb95f1891a2efb747ed10761cdfe1cde12d81a","last_reissued_at":"2026-07-05T08:28:05.090884Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:05.090884Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond Traditional Benchmarks: Analyzing Behaviors of Open LLMs on Data-to-Text Generation","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ond\\v{r}ej Du\\v{s}ek, Zden\\v{e}k Kasner","submitted_at":"2024-01-18T18:15:46Z","abstract_excerpt":"We analyze the behaviors of open large language models (LLMs) on the task of data-to-text (D2T) generation, i.e., generating coherent and relevant text from structured data. To avoid the issue of LLM training data contamination with standard benchmarks, we design Quintd - a tool for collecting novel structured data records from public APIs. We find that open LLMs (Llama 2, Mistral, and Zephyr) can generate fluent and coherent texts in zero-shot settings from data in common formats collected with Quintd. However, we show that the semantic accuracy of the outputs is a major issue: both according"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.10186","kind":"arxiv","version":3},"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/2401.10186/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":"2401.10186","created_at":"2026-07-05T08:28:05.090934+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.10186v3","created_at":"2026-07-05T08:28:05.090934+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.10186","created_at":"2026-07-05T08:28:05.090934+00:00"},{"alias_kind":"pith_short_12","alias_value":"KQNPUH37OGA5","created_at":"2026-07-05T08:28:05.090934+00:00"},{"alias_kind":"pith_short_16","alias_value":"KQNPUH37OGA5FVTF","created_at":"2026-07-05T08:28:05.090934+00:00"},{"alias_kind":"pith_short_8","alias_value":"KQNPUH37","created_at":"2026-07-05T08:28:05.090934+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.19203","citing_title":"An Extensive Evaluation of Factual Consistency in Large Language Models for Data-to-Text Generation","ref_index":19,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G","json":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G.json","graph_json":"https://pith.science/api/pith-number/KQNPUH37OGA5FVTFIJM7564V6G/graph.json","events_json":"https://pith.science/api/pith-number/KQNPUH37OGA5FVTFIJM7564V6G/events.json","paper":"https://pith.science/paper/KQNPUH37"},"agent_actions":{"view_html":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G","download_json":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G.json","view_paper":"https://pith.science/paper/KQNPUH37","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.10186&json=true","fetch_graph":"https://pith.science/api/pith-number/KQNPUH37OGA5FVTFIJM7564V6G/graph.json","fetch_events":"https://pith.science/api/pith-number/KQNPUH37OGA5FVTFIJM7564V6G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G/action/storage_attestation","attest_author":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G/action/author_attestation","sign_citation":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G/action/citation_signature","submit_replication":"https://pith.science/pith/KQNPUH37OGA5FVTFIJM7564V6G/action/replication_record"}},"created_at":"2026-07-05T08:28:05.090934+00:00","updated_at":"2026-07-05T08:28:05.090934+00:00"}