{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:P7VNWM6BU6YIND4JFRYYASXHPL","short_pith_number":"pith:P7VNWM6B","schema_version":"1.0","canonical_sha256":"7feadb33c1a7b0868f892c71804ae77afcf9da4f6cf3dac039874b611df65361","source":{"kind":"arxiv","id":"2312.06742","version":2},"attestation_state":"computed","paper":{"title":"Honeybee: Locality-enhanced Projector for Multimodal LLM","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Byungseok Roh, Jonghwan Mun, Junbum Cha, Wooyoung Kang","submitted_at":"2023-12-11T18:59:06Z","abstract_excerpt":"In Multimodal Large Language Models (MLLMs), a visual projector plays a crucial role in bridging pre-trained vision encoders with LLMs, enabling profound visual understanding while harnessing the LLMs' robust capabilities. Despite the importance of the visual projector, it has been relatively less explored. In this study, we first identify two essential projector properties: (i) flexibility in managing the number of visual tokens, crucial for MLLMs' overall efficiency, and (ii) preservation of local context from visual features, vital for spatial understanding. Based on these findings, we prop"},"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":"2312.06742","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-12-11T18:59:06Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"f07b413fae7416e52a7b13530a058496ad50fcb0f089e93257768289efaf21e1","abstract_canon_sha256":"3339bb026c328f40d780a25d6cd7720439715812ad6bbe49bc51f0a110391809"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:02:49.172902Z","signature_b64":"/q77tLmYsvtcN/NSRbtQ32W0XwC1d+PgGhmO5Bx3y4WuGfKfTnOO8lc4/JbfdgtlUTMxpWT1lHL87WKSy7MtDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7feadb33c1a7b0868f892c71804ae77afcf9da4f6cf3dac039874b611df65361","last_reissued_at":"2026-07-05T08:02:49.172423Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:02:49.172423Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Honeybee: Locality-enhanced Projector for Multimodal LLM","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Byungseok Roh, Jonghwan Mun, Junbum Cha, Wooyoung Kang","submitted_at":"2023-12-11T18:59:06Z","abstract_excerpt":"In Multimodal Large Language Models (MLLMs), a visual projector plays a crucial role in bridging pre-trained vision encoders with LLMs, enabling profound visual understanding while harnessing the LLMs' robust capabilities. Despite the importance of the visual projector, it has been relatively less explored. In this study, we first identify two essential projector properties: (i) flexibility in managing the number of visual tokens, crucial for MLLMs' overall efficiency, and (ii) preservation of local context from visual features, vital for spatial understanding. Based on these findings, we prop"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.06742","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/2312.06742/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":"2312.06742","created_at":"2026-07-05T08:02:49.172481+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.06742v2","created_at":"2026-07-05T08:02:49.172481+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.06742","created_at":"2026-07-05T08:02:49.172481+00:00"},{"alias_kind":"pith_short_12","alias_value":"P7VNWM6BU6YI","created_at":"2026-07-05T08:02:49.172481+00:00"},{"alias_kind":"pith_short_16","alias_value":"P7VNWM6BU6YIND4J","created_at":"2026-07-05T08:02:49.172481+00:00"},{"alias_kind":"pith_short_8","alias_value":"P7VNWM6B","created_at":"2026-07-05T08:02:49.172481+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21734","citing_title":"HPP: Hierarchical Programmatic Probing for Long Video Understanding by Decoupling Perception and Reasoning","ref_index":246,"is_internal_anchor":false},{"citing_arxiv_id":"2402.03766","citing_title":"MobileVLM V2: Faster and Stronger Baseline for Vision Language Model","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2402.11411","citing_title":"Aligning Modalities in Vision Large Language Models via Preference Fine-tuning","ref_index":145,"is_internal_anchor":false},{"citing_arxiv_id":"2403.09611","citing_title":"MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2401.15947","citing_title":"MoE-LLaVA: Mixture of Experts for Large Vision-Language Models","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2407.07726","citing_title":"PaliGemma: A versatile 3B VLM for transfer","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL","json":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL.json","graph_json":"https://pith.science/api/pith-number/P7VNWM6BU6YIND4JFRYYASXHPL/graph.json","events_json":"https://pith.science/api/pith-number/P7VNWM6BU6YIND4JFRYYASXHPL/events.json","paper":"https://pith.science/paper/P7VNWM6B"},"agent_actions":{"view_html":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL","download_json":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL.json","view_paper":"https://pith.science/paper/P7VNWM6B","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.06742&json=true","fetch_graph":"https://pith.science/api/pith-number/P7VNWM6BU6YIND4JFRYYASXHPL/graph.json","fetch_events":"https://pith.science/api/pith-number/P7VNWM6BU6YIND4JFRYYASXHPL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL/action/storage_attestation","attest_author":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL/action/author_attestation","sign_citation":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL/action/citation_signature","submit_replication":"https://pith.science/pith/P7VNWM6BU6YIND4JFRYYASXHPL/action/replication_record"}},"created_at":"2026-07-05T08:02:49.172481+00:00","updated_at":"2026-07-05T08:02:49.172481+00:00"}