{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:TMM7B6CXLYCFKJHFSEOJFT3T2W","short_pith_number":"pith:TMM7B6CX","schema_version":"1.0","canonical_sha256":"9b19f0f8575e045524e5911c92cf73d5bcb858e0dfff1ca70de2f67becb53c07","source":{"kind":"arxiv","id":"2411.10557","version":3},"attestation_state":"computed","paper":{"title":"MLAN: Language-Based Instruction Tuning Preserves and Transfers Knowledge in Multimodal Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Beliz Gunel, Chenguang Wang, Dawn Song, Jianhong Tu, Lingjuan Lyu, Michael Bendersky, Nicholas Crispino, Ruoxi Jia, Xin Liu, Zhuohao Ni, Zihao Yu","submitted_at":"2024-11-15T20:09:59Z","abstract_excerpt":"We present a novel visual instruction tuning strategy to improve the zero-shot task generalization of multimodal large language models by building a firm text-only knowledge base. Existing work lacks sufficient experimentation on the importance of each modality in the instruction tuning stage, often using a majority of vision-language data while keeping text-only data limited and fixing mixtures of modalities. By incorporating diverse text-only data in the visual instruction tuning stage, we vary vision-language data in various controlled experiments to investigate the importance of modality i"},"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":"2411.10557","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-15T20:09:59Z","cross_cats_sorted":[],"title_canon_sha256":"3eaa7bfe00724170f4fb5820718314eb8805d9f5a8e7d1ab591df58b316e4ea1","abstract_canon_sha256":"96336dda458b705570a96262cba2e0555047ab1794e55bbd21a174837900473c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:28:32.393508Z","signature_b64":"YgR+CjYjjc2CWZSe1TVNbJKMrxHFL5pKvR+4vZKiKUjKhHAMnanT+Aaj15d0oNOhi5SuMUtbrirtMv3WuEXrCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9b19f0f8575e045524e5911c92cf73d5bcb858e0dfff1ca70de2f67becb53c07","last_reissued_at":"2026-07-05T11:28:32.393002Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:28:32.393002Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MLAN: Language-Based Instruction Tuning Preserves and Transfers Knowledge in Multimodal Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Beliz Gunel, Chenguang Wang, Dawn Song, Jianhong Tu, Lingjuan Lyu, Michael Bendersky, Nicholas Crispino, Ruoxi Jia, Xin Liu, Zhuohao Ni, Zihao Yu","submitted_at":"2024-11-15T20:09:59Z","abstract_excerpt":"We present a novel visual instruction tuning strategy to improve the zero-shot task generalization of multimodal large language models by building a firm text-only knowledge base. Existing work lacks sufficient experimentation on the importance of each modality in the instruction tuning stage, often using a majority of vision-language data while keeping text-only data limited and fixing mixtures of modalities. By incorporating diverse text-only data in the visual instruction tuning stage, we vary vision-language data in various controlled experiments to investigate the importance of modality i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.10557","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/2411.10557/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":"2411.10557","created_at":"2026-07-05T11:28:32.393063+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.10557v3","created_at":"2026-07-05T11:28:32.393063+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.10557","created_at":"2026-07-05T11:28:32.393063+00:00"},{"alias_kind":"pith_short_12","alias_value":"TMM7B6CXLYCF","created_at":"2026-07-05T11:28:32.393063+00:00"},{"alias_kind":"pith_short_16","alias_value":"TMM7B6CXLYCFKJHF","created_at":"2026-07-05T11:28:32.393063+00:00"},{"alias_kind":"pith_short_8","alias_value":"TMM7B6CX","created_at":"2026-07-05T11:28:32.393063+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.14029","citing_title":"POINTS-Seeker: An Open Recipe for Multimodal Search Agents with Visual Memory Management","ref_index":42,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W","json":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W.json","graph_json":"https://pith.science/api/pith-number/TMM7B6CXLYCFKJHFSEOJFT3T2W/graph.json","events_json":"https://pith.science/api/pith-number/TMM7B6CXLYCFKJHFSEOJFT3T2W/events.json","paper":"https://pith.science/paper/TMM7B6CX"},"agent_actions":{"view_html":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W","download_json":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W.json","view_paper":"https://pith.science/paper/TMM7B6CX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.10557&json=true","fetch_graph":"https://pith.science/api/pith-number/TMM7B6CXLYCFKJHFSEOJFT3T2W/graph.json","fetch_events":"https://pith.science/api/pith-number/TMM7B6CXLYCFKJHFSEOJFT3T2W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W/action/storage_attestation","attest_author":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W/action/author_attestation","sign_citation":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W/action/citation_signature","submit_replication":"https://pith.science/pith/TMM7B6CXLYCFKJHFSEOJFT3T2W/action/replication_record"}},"created_at":"2026-07-05T11:28:32.393063+00:00","updated_at":"2026-07-05T11:28:32.393063+00:00"}