{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:LWUXMX44I253TBOD5RDNYADGER","short_pith_number":"pith:LWUXMX44","schema_version":"1.0","canonical_sha256":"5da9765f9c46bbb985c3ec46dc00662475e537256ba388a57de951bba9dbe0bd","source":{"kind":"arxiv","id":"2412.14642","version":4},"attestation_state":"computed","paper":{"title":"Speak-to-Structure: Evaluating LLMs in Open-domain Natural Language-Driven Molecule Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Changmeng Zheng, Dongzhan Zhou, Jiatong Li, Junxian Li, Qing Li, Weida Wang, Xiao-Yong Wei, Yatao Bian, Yunqing Liu","submitted_at":"2024-12-19T08:51:16Z","abstract_excerpt":"Recently, Large Language Models (LLMs) have demonstrated great potential in natural language-driven molecule discovery. However, existing datasets and benchmarks for molecule-text alignment are predominantly built on one-to-one mappings, measuring LLMs' ability to retrieve a single, pre-defined answer, rather than their creative potential to generate diverse, yet equally valid, molecular candidates. To address this critical gap, we propose Speak-to-Structure (S^2-Bench), the first benchmark to evaluate LLMs in open-domain natural language-driven molecule generation. S^2-Bench is specifically d"},"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":"2412.14642","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-12-19T08:51:16Z","cross_cats_sorted":[],"title_canon_sha256":"cdc6c637e421000c6d4fa0fbc6d251086ff2f7763d9753924b1a2c1feefe8bff","abstract_canon_sha256":"d9906cba0e008c1f18418fb9045727f99a090e85ce2c2368fb800a952e1db768"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:02:06.129197Z","signature_b64":"F41oeoEcuk+jEUBuzb7+kbEpxFwnj/Kh2O+N+YSVK2+bKaqCQpaggUuI88IU6wgu9YvK2Wqxnr90leuc6tN7AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5da9765f9c46bbb985c3ec46dc00662475e537256ba388a57de951bba9dbe0bd","last_reissued_at":"2026-05-25T02:02:06.128255Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:02:06.128255Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Speak-to-Structure: Evaluating LLMs in Open-domain Natural Language-Driven Molecule Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Changmeng Zheng, Dongzhan Zhou, Jiatong Li, Junxian Li, Qing Li, Weida Wang, Xiao-Yong Wei, Yatao Bian, Yunqing Liu","submitted_at":"2024-12-19T08:51:16Z","abstract_excerpt":"Recently, Large Language Models (LLMs) have demonstrated great potential in natural language-driven molecule discovery. However, existing datasets and benchmarks for molecule-text alignment are predominantly built on one-to-one mappings, measuring LLMs' ability to retrieve a single, pre-defined answer, rather than their creative potential to generate diverse, yet equally valid, molecular candidates. To address this critical gap, we propose Speak-to-Structure (S^2-Bench), the first benchmark to evaluate LLMs in open-domain natural language-driven molecule generation. S^2-Bench is specifically d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.14642","kind":"arxiv","version":4},"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/2412.14642/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":"2412.14642","created_at":"2026-05-25T02:02:06.128409+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.14642v4","created_at":"2026-05-25T02:02:06.128409+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.14642","created_at":"2026-05-25T02:02:06.128409+00:00"},{"alias_kind":"pith_short_12","alias_value":"LWUXMX44I253","created_at":"2026-05-25T02:02:06.128409+00:00"},{"alias_kind":"pith_short_16","alias_value":"LWUXMX44I253TBOD","created_at":"2026-05-25T02:02:06.128409+00:00"},{"alias_kind":"pith_short_8","alias_value":"LWUXMX44","created_at":"2026-05-25T02:02:06.128409+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":4,"sample":[{"citing_arxiv_id":"2605.16480","citing_title":"MoleCode unlocks structural intelligence in large language models","ref_index":58,"is_internal_anchor":true},{"citing_arxiv_id":"2604.18031","citing_title":"How Creative Are Large Language Models in Generating Molecules?","ref_index":42,"is_internal_anchor":true},{"citing_arxiv_id":"2604.14116","citing_title":"TREX: Automating LLM Fine-tuning via Agent-Driven Tree-based Exploration","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2605.02351","citing_title":"MolViBench: Evaluating LLMs on Molecular Vibe Coding","ref_index":7,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER","json":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER.json","graph_json":"https://pith.science/api/pith-number/LWUXMX44I253TBOD5RDNYADGER/graph.json","events_json":"https://pith.science/api/pith-number/LWUXMX44I253TBOD5RDNYADGER/events.json","paper":"https://pith.science/paper/LWUXMX44"},"agent_actions":{"view_html":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER","download_json":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER.json","view_paper":"https://pith.science/paper/LWUXMX44","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.14642&json=true","fetch_graph":"https://pith.science/api/pith-number/LWUXMX44I253TBOD5RDNYADGER/graph.json","fetch_events":"https://pith.science/api/pith-number/LWUXMX44I253TBOD5RDNYADGER/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER/action/storage_attestation","attest_author":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER/action/author_attestation","sign_citation":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER/action/citation_signature","submit_replication":"https://pith.science/pith/LWUXMX44I253TBOD5RDNYADGER/action/replication_record"}},"created_at":"2026-05-25T02:02:06.128409+00:00","updated_at":"2026-05-25T02:02:06.128409+00:00"}