{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KT77VXKDPWBL4WL6IZHGHKYPW7","short_pith_number":"pith:KT77VXKD","schema_version":"1.0","canonical_sha256":"54fffadd437d82be597e464e63ab0fb7d79861743e2ee28834492a00f04e9546","source":{"kind":"arxiv","id":"2504.04633","version":3},"attestation_state":"computed","paper":{"title":"M$^2$IV: Towards Efficient and Fine-grained Multimodal In-Context Learning via Representation Engineering","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Hongyang He, Qisen Cheng, Ruixiang Tang, Tianyang Wang, Xiang Fu, Xi Xiao, Yanshu Li, Yi Cao","submitted_at":"2025-04-06T22:02:21Z","abstract_excerpt":"Multimodal in-context learning (ICL) equips Large Vision-language Models (LVLMs) with the ability to adapt to new tasks via multiple user-provided demonstrations, without requiring any model parameter updates. However, its effectiveness is constrained by the token-intensive nature of multimodal inputs and the complexity of cross-modal few-shot reasoning, which together hinder LVLMs from extracting useful patterns from demonstrations. To address these challenges, we propose \\textbf{M$^2$IV}, a novel representation engineering approach that replaces explicit token-level demonstrations with a set"},"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":"2504.04633","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2025-04-06T22:02:21Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"88a1c695cb079b8af9f8b572064e6215a6aa39da693c7cf71d20b754219a301e","abstract_canon_sha256":"e0c37b7be6755e27983ee81192ce1e6fa317d91b88d8ea018849f66ed16d75a7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:59:12.742795Z","signature_b64":"9UDxroFdL4ZDVbwFT0fBiGEEwLeoF+6c82Yfhe8e8I7INAdZKXY4elcg6EkcYTZCDkUEYrqBDfTLKGqAPvyWBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"54fffadd437d82be597e464e63ab0fb7d79861743e2ee28834492a00f04e9546","last_reissued_at":"2026-07-05T11:59:12.742351Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:59:12.742351Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"M$^2$IV: Towards Efficient and Fine-grained Multimodal In-Context Learning via Representation Engineering","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Hongyang He, Qisen Cheng, Ruixiang Tang, Tianyang Wang, Xiang Fu, Xi Xiao, Yanshu Li, Yi Cao","submitted_at":"2025-04-06T22:02:21Z","abstract_excerpt":"Multimodal in-context learning (ICL) equips Large Vision-language Models (LVLMs) with the ability to adapt to new tasks via multiple user-provided demonstrations, without requiring any model parameter updates. However, its effectiveness is constrained by the token-intensive nature of multimodal inputs and the complexity of cross-modal few-shot reasoning, which together hinder LVLMs from extracting useful patterns from demonstrations. To address these challenges, we propose \\textbf{M$^2$IV}, a novel representation engineering approach that replaces explicit token-level demonstrations with a set"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.04633","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/2504.04633/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":"2504.04633","created_at":"2026-07-05T11:59:12.742404+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.04633v3","created_at":"2026-07-05T11:59:12.742404+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.04633","created_at":"2026-07-05T11:59:12.742404+00:00"},{"alias_kind":"pith_short_12","alias_value":"KT77VXKDPWBL","created_at":"2026-07-05T11:59:12.742404+00:00"},{"alias_kind":"pith_short_16","alias_value":"KT77VXKDPWBL4WL6","created_at":"2026-07-05T11:59:12.742404+00:00"},{"alias_kind":"pith_short_8","alias_value":"KT77VXKD","created_at":"2026-07-05T11:59:12.742404+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.11103","citing_title":"ActorMind: Emulating Human Actor Reasoning for Speech Role-Playing","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02378","citing_title":"Enhancing Multimodal In-Context Learning via Inductive-Deductive Reasoning","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7","json":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7.json","graph_json":"https://pith.science/api/pith-number/KT77VXKDPWBL4WL6IZHGHKYPW7/graph.json","events_json":"https://pith.science/api/pith-number/KT77VXKDPWBL4WL6IZHGHKYPW7/events.json","paper":"https://pith.science/paper/KT77VXKD"},"agent_actions":{"view_html":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7","download_json":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7.json","view_paper":"https://pith.science/paper/KT77VXKD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.04633&json=true","fetch_graph":"https://pith.science/api/pith-number/KT77VXKDPWBL4WL6IZHGHKYPW7/graph.json","fetch_events":"https://pith.science/api/pith-number/KT77VXKDPWBL4WL6IZHGHKYPW7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7/action/storage_attestation","attest_author":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7/action/author_attestation","sign_citation":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7/action/citation_signature","submit_replication":"https://pith.science/pith/KT77VXKDPWBL4WL6IZHGHKYPW7/action/replication_record"}},"created_at":"2026-07-05T11:59:12.742404+00:00","updated_at":"2026-07-05T11:59:12.742404+00:00"}