{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QYS47AKMLNKS36QXO6ENETIVZU","short_pith_number":"pith:QYS47AKM","schema_version":"1.0","canonical_sha256":"8625cf814c5b552dfa177788d24d15cd121358a1cdd0ab504917059fc88023d4","source":{"kind":"arxiv","id":"2401.16160","version":2},"attestation_state":"computed","paper":{"title":"LLaVA-MoLE: Sparse Mixture of LoRA Experts for Mitigating Data Conflicts in Instruction Finetuning MLLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Lin Ma, Shaoxiang Chen, Zequn Jie","submitted_at":"2024-01-29T13:48:36Z","abstract_excerpt":"Instruction finetuning on a variety of image-text instruction data is the key to obtaining a versatile Multimodal Large Language Model (MLLM), and different configurations of the instruction data can lead to finetuned models with different capabilities. However, we have discovered that data conflicts are inevitable when mixing instruction data from distinct domains, which can result in performance drops for tasks of a specific domain. To address this issue, we propose to apply an efficient Mixture of Experts (MoE) design, which is a sparse Mixture of LoRA Experts (MoLE) for instruction finetun"},"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":"2401.16160","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-01-29T13:48:36Z","cross_cats_sorted":[],"title_canon_sha256":"9d25ef0f0e465efcc02e531504574a0b780d89af1b3e94d8c16ee0da02f755e8","abstract_canon_sha256":"0780ebc4f3652ddf7b7dfc08640eea5e2651d15fd49a67b570be6d07760b15bc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:39:15.252151Z","signature_b64":"/nMb3rFnFCrliY0ubSKQQqfODqnaKbQRS1x9KBedT2zTCyKLzFn9On7+ICUAkfvp1+RSo8twWfnrnsf4HbLqAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8625cf814c5b552dfa177788d24d15cd121358a1cdd0ab504917059fc88023d4","last_reissued_at":"2026-07-05T07:39:15.251744Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:39:15.251744Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLaVA-MoLE: Sparse Mixture of LoRA Experts for Mitigating Data Conflicts in Instruction Finetuning MLLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Lin Ma, Shaoxiang Chen, Zequn Jie","submitted_at":"2024-01-29T13:48:36Z","abstract_excerpt":"Instruction finetuning on a variety of image-text instruction data is the key to obtaining a versatile Multimodal Large Language Model (MLLM), and different configurations of the instruction data can lead to finetuned models with different capabilities. However, we have discovered that data conflicts are inevitable when mixing instruction data from distinct domains, which can result in performance drops for tasks of a specific domain. To address this issue, we propose to apply an efficient Mixture of Experts (MoE) design, which is a sparse Mixture of LoRA Experts (MoLE) for instruction finetun"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.16160","kind":"arxiv","version":2},"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/2401.16160/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":"2401.16160","created_at":"2026-07-05T07:39:15.251811+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.16160v2","created_at":"2026-07-05T07:39:15.251811+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.16160","created_at":"2026-07-05T07:39:15.251811+00:00"},{"alias_kind":"pith_short_12","alias_value":"QYS47AKMLNKS","created_at":"2026-07-05T07:39:15.251811+00:00"},{"alias_kind":"pith_short_16","alias_value":"QYS47AKMLNKS36QX","created_at":"2026-07-05T07:39:15.251811+00:00"},{"alias_kind":"pith_short_8","alias_value":"QYS47AKM","created_at":"2026-07-05T07:39:15.251811+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21645","citing_title":"Behavioral and Representational Evidence of Binomial Ordering Preferences in Large Language Models","ref_index":125,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20970","citing_title":"CogniRoute: Learning to Route Social Evidence in Omni-Modal Models","ref_index":134,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10488","citing_title":"5% > 100%: Flatness Preference is All You Need for Multimodal Parameter-Efficient Fine-Tuning","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04349","citing_title":"MorphoQuant: Modality-Aware Quantization for Omni-modal Large Language Models","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00275","citing_title":"Hyperbolic and Evidence-Prioritized Experts for Large Vision-Language Models","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2506.21035","citing_title":"Little by Little: Continual Learning via Incremental Mixture of Rank-1 Associative Memory Experts","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18192","citing_title":"View-Aware Semantic Alignment for Aerial-Ground Person Re-Identification","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2507.00029","citing_title":"LoRA-Mixer: Coordinate Modular LoRA Experts Through Serial Attention Routing","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23996","citing_title":"SMoES: Soft Modality-Guided Expert Specialization in MoE-VLMs","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU","json":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU.json","graph_json":"https://pith.science/api/pith-number/QYS47AKMLNKS36QXO6ENETIVZU/graph.json","events_json":"https://pith.science/api/pith-number/QYS47AKMLNKS36QXO6ENETIVZU/events.json","paper":"https://pith.science/paper/QYS47AKM"},"agent_actions":{"view_html":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU","download_json":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU.json","view_paper":"https://pith.science/paper/QYS47AKM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.16160&json=true","fetch_graph":"https://pith.science/api/pith-number/QYS47AKMLNKS36QXO6ENETIVZU/graph.json","fetch_events":"https://pith.science/api/pith-number/QYS47AKMLNKS36QXO6ENETIVZU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU/action/storage_attestation","attest_author":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU/action/author_attestation","sign_citation":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU/action/citation_signature","submit_replication":"https://pith.science/pith/QYS47AKMLNKS36QXO6ENETIVZU/action/replication_record"}},"created_at":"2026-07-05T07:39:15.251811+00:00","updated_at":"2026-07-05T07:39:15.251811+00:00"}