{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RGJ7MLSK3MFOERYPRPLR3C6IX4","short_pith_number":"pith:RGJ7MLSK","schema_version":"1.0","canonical_sha256":"8993f62e4adb0ae2470f8bd71d8bc8bf2dbb65d0d4ac4113ee07120a21628766","source":{"kind":"arxiv","id":"2408.06610","version":1},"attestation_state":"computed","paper":{"title":"CROME: Cross-Modal Adapters for Efficient Multimodal LLM","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Sayna Ebrahimi, Sercan O. Arik, Tejas Nama, Tomas Pfister","submitted_at":"2024-08-13T03:45:11Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) demonstrate remarkable image-language capabilities, but their widespread use faces challenges in cost-effective training and adaptation. Existing approaches often necessitate expensive language model retraining and limited adaptability. Additionally, the current focus on zero-shot performance improvements offers insufficient guidance for task-specific tuning. We propose CROME, an efficient vision-language instruction tuning framework. It features a novel gated cross-modal adapter that effectively combines visual and textual representations prior to inpu"},"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":"2408.06610","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-08-13T03:45:11Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"5a56854ab0bd4f49ad3c4e5ccee8bd6c740a2bc24561b8284c06b5183ffed04a","abstract_canon_sha256":"c373871f1315eeecb3d6d96f2ae6052041c7d8fe9d44389c4f7d60adfcac880f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:54:59.710949Z","signature_b64":"Cm/MiBBYS/lqaKFg0t6iIEyBk3jb2vnn2MeTxf/7fYPQWheurL2GjQh7XSykYjMRJWbEcHcTbCmT22s0OGnfDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8993f62e4adb0ae2470f8bd71d8bc8bf2dbb65d0d4ac4113ee07120a21628766","last_reissued_at":"2026-07-05T08:54:59.710520Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:54:59.710520Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CROME: Cross-Modal Adapters for Efficient Multimodal LLM","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Sayna Ebrahimi, Sercan O. Arik, Tejas Nama, Tomas Pfister","submitted_at":"2024-08-13T03:45:11Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) demonstrate remarkable image-language capabilities, but their widespread use faces challenges in cost-effective training and adaptation. Existing approaches often necessitate expensive language model retraining and limited adaptability. Additionally, the current focus on zero-shot performance improvements offers insufficient guidance for task-specific tuning. We propose CROME, an efficient vision-language instruction tuning framework. It features a novel gated cross-modal adapter that effectively combines visual and textual representations prior to inpu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.06610","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/2408.06610/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":"2408.06610","created_at":"2026-07-05T08:54:59.710576+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.06610v1","created_at":"2026-07-05T08:54:59.710576+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.06610","created_at":"2026-07-05T08:54:59.710576+00:00"},{"alias_kind":"pith_short_12","alias_value":"RGJ7MLSK3MFO","created_at":"2026-07-05T08:54:59.710576+00:00"},{"alias_kind":"pith_short_16","alias_value":"RGJ7MLSK3MFOERYP","created_at":"2026-07-05T08:54:59.710576+00:00"},{"alias_kind":"pith_short_8","alias_value":"RGJ7MLSK","created_at":"2026-07-05T08:54:59.710576+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.24675","citing_title":"VaaWIT: Visual-Aware Adaptation of Large Language Models for Multilingual Web Image Translation","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4","json":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4.json","graph_json":"https://pith.science/api/pith-number/RGJ7MLSK3MFOERYPRPLR3C6IX4/graph.json","events_json":"https://pith.science/api/pith-number/RGJ7MLSK3MFOERYPRPLR3C6IX4/events.json","paper":"https://pith.science/paper/RGJ7MLSK"},"agent_actions":{"view_html":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4","download_json":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4.json","view_paper":"https://pith.science/paper/RGJ7MLSK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.06610&json=true","fetch_graph":"https://pith.science/api/pith-number/RGJ7MLSK3MFOERYPRPLR3C6IX4/graph.json","fetch_events":"https://pith.science/api/pith-number/RGJ7MLSK3MFOERYPRPLR3C6IX4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4/action/storage_attestation","attest_author":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4/action/author_attestation","sign_citation":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4/action/citation_signature","submit_replication":"https://pith.science/pith/RGJ7MLSK3MFOERYPRPLR3C6IX4/action/replication_record"}},"created_at":"2026-07-05T08:54:59.710576+00:00","updated_at":"2026-07-05T08:54:59.710576+00:00"}