{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5ECPK2PJEHTR3OGE7VCZGBVC74","short_pith_number":"pith:5ECPK2PJ","schema_version":"1.0","canonical_sha256":"e904f569e921e71db8c4fd459306a2ff1962c3b8d876cc578ef3c0599d10f676","source":{"kind":"arxiv","id":"2412.02368","version":1},"attestation_state":"computed","paper":{"title":"ScImage: How Good Are Multimodal Large Language Models at Scientific Text-to-Image Generation?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.AI","authors_text":"Christoph Leiter, Fahimeh Moafian, Jonas Belouadi, Leixin Zhang, Simone Paolo Ponzetto, Steffen Eger, Weihe Zhai, Yinjie Cheng, Zhixue Zhao","submitted_at":"2024-12-03T10:52:06Z","abstract_excerpt":"Multimodal large language models (LLMs) have demonstrated impressive capabilities in generating high-quality images from textual instructions. However, their performance in generating scientific images--a critical application for accelerating scientific progress--remains underexplored. In this work, we address this gap by introducing ScImage, a benchmark designed to evaluate the multimodal capabilities of LLMs in generating scientific images from textual descriptions. ScImage assesses three key dimensions of understanding: spatial, numeric, and attribute comprehension, as well as their combina"},"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":"2412.02368","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-12-03T10:52:06Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"3236801c8ad08315d978b990f4f5374dde23277a18d0a04c43dd90fd6bbbffd6","abstract_canon_sha256":"95d22942ec444229f6ff5d2adb163f3cb295883cfb5364a831a8bf2db64a3035"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:43:48.785320Z","signature_b64":"VKoAVcQ0Wo2wFhix37KYgYaZMvMJIomrng7igbp3l1weuCX4iPvUhNYxtDAkOT37p5ZhmCHtEgtP0ry//GPqDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e904f569e921e71db8c4fd459306a2ff1962c3b8d876cc578ef3c0599d10f676","last_reissued_at":"2026-07-05T09:43:48.784916Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:43:48.784916Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ScImage: How Good Are Multimodal Large Language Models at Scientific Text-to-Image Generation?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.AI","authors_text":"Christoph Leiter, Fahimeh Moafian, Jonas Belouadi, Leixin Zhang, Simone Paolo Ponzetto, Steffen Eger, Weihe Zhai, Yinjie Cheng, Zhixue Zhao","submitted_at":"2024-12-03T10:52:06Z","abstract_excerpt":"Multimodal large language models (LLMs) have demonstrated impressive capabilities in generating high-quality images from textual instructions. However, their performance in generating scientific images--a critical application for accelerating scientific progress--remains underexplored. In this work, we address this gap by introducing ScImage, a benchmark designed to evaluate the multimodal capabilities of LLMs in generating scientific images from textual descriptions. ScImage assesses three key dimensions of understanding: spatial, numeric, and attribute comprehension, as well as their combina"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.02368","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/2412.02368/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":"2412.02368","created_at":"2026-07-05T09:43:48.784966+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.02368v1","created_at":"2026-07-05T09:43:48.784966+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.02368","created_at":"2026-07-05T09:43:48.784966+00:00"},{"alias_kind":"pith_short_12","alias_value":"5ECPK2PJEHTR","created_at":"2026-07-05T09:43:48.784966+00:00"},{"alias_kind":"pith_short_16","alias_value":"5ECPK2PJEHTR3OGE","created_at":"2026-07-05T09:43:48.784966+00:00"},{"alias_kind":"pith_short_8","alias_value":"5ECPK2PJ","created_at":"2026-07-05T09:43:48.784966+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.30124","citing_title":"SciIR: A Large-scale Training Dataset and Benchmark for Scientific Image Reasoning Generation","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01168","citing_title":"Quantifying and Predicting Disagreement in Graded Human Ratings","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18069","citing_title":"Modeling Human Perspectives with Socio-Demographic Representations","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74","json":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74.json","graph_json":"https://pith.science/api/pith-number/5ECPK2PJEHTR3OGE7VCZGBVC74/graph.json","events_json":"https://pith.science/api/pith-number/5ECPK2PJEHTR3OGE7VCZGBVC74/events.json","paper":"https://pith.science/paper/5ECPK2PJ"},"agent_actions":{"view_html":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74","download_json":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74.json","view_paper":"https://pith.science/paper/5ECPK2PJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.02368&json=true","fetch_graph":"https://pith.science/api/pith-number/5ECPK2PJEHTR3OGE7VCZGBVC74/graph.json","fetch_events":"https://pith.science/api/pith-number/5ECPK2PJEHTR3OGE7VCZGBVC74/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74/action/storage_attestation","attest_author":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74/action/author_attestation","sign_citation":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74/action/citation_signature","submit_replication":"https://pith.science/pith/5ECPK2PJEHTR3OGE7VCZGBVC74/action/replication_record"}},"created_at":"2026-07-05T09:43:48.784966+00:00","updated_at":"2026-07-05T09:43:48.784966+00:00"}