{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:C63I6HX6STBY6BESHFP64WDARH","short_pith_number":"pith:C63I6HX6","schema_version":"1.0","canonical_sha256":"17b68f1efe94c38f0492395fee586089d95b69d2c32612e7f7fe6501f1575f85","source":{"kind":"arxiv","id":"2504.04024","version":1},"attestation_state":"computed","paper":{"title":"Window Token Concatenation for Efficient Visual Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Botao Ye, Huan Liu, Tianlong Chen, Wentao Bao, Yifan Li, Yu Kong, Zhen Tan","submitted_at":"2025-04-05T02:32:58Z","abstract_excerpt":"To effectively reduce the visual tokens in Visual Large Language Models (VLLMs), we propose a novel approach called Window Token Concatenation (WiCo). Specifically, we employ a sliding window to concatenate spatially adjacent visual tokens. However, directly concatenating these tokens may group diverse tokens into one, and thus obscure some fine details. To address this challenge, we propose fine-tuning the last few layers of the vision encoder to adaptively adjust the visual tokens, encouraging that those within the same window exhibit similar features. To further enhance the performance on f"},"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":"2504.04024","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-04-05T02:32:58Z","cross_cats_sorted":[],"title_canon_sha256":"64a616526519c61e08e98477703035cb472b1f18254a03a65b9ade3647c8c5a6","abstract_canon_sha256":"50f2e1cc73e8ddf8a00c16ba0440ad2b548a71c323c737e0c5b94fd630a18cb5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:44:53.104254Z","signature_b64":"JgpDUoh1TiR+cG0tIlIO77YfRMMqJ1B1j0UW4bhi1XNJexEak6V+BQdlPyVCAQU+CdZ600u0rqYNXZ0STjtcBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"17b68f1efe94c38f0492395fee586089d95b69d2c32612e7f7fe6501f1575f85","last_reissued_at":"2026-07-05T10:44:53.103765Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:44:53.103765Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Window Token Concatenation for Efficient Visual Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Botao Ye, Huan Liu, Tianlong Chen, Wentao Bao, Yifan Li, Yu Kong, Zhen Tan","submitted_at":"2025-04-05T02:32:58Z","abstract_excerpt":"To effectively reduce the visual tokens in Visual Large Language Models (VLLMs), we propose a novel approach called Window Token Concatenation (WiCo). Specifically, we employ a sliding window to concatenate spatially adjacent visual tokens. However, directly concatenating these tokens may group diverse tokens into one, and thus obscure some fine details. To address this challenge, we propose fine-tuning the last few layers of the vision encoder to adaptively adjust the visual tokens, encouraging that those within the same window exhibit similar features. To further enhance the performance on f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.04024","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/2504.04024/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":"2504.04024","created_at":"2026-07-05T10:44:53.103824+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.04024v1","created_at":"2026-07-05T10:44:53.103824+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.04024","created_at":"2026-07-05T10:44:53.103824+00:00"},{"alias_kind":"pith_short_12","alias_value":"C63I6HX6STBY","created_at":"2026-07-05T10:44:53.103824+00:00"},{"alias_kind":"pith_short_16","alias_value":"C63I6HX6STBY6BES","created_at":"2026-07-05T10:44:53.103824+00:00"},{"alias_kind":"pith_short_8","alias_value":"C63I6HX6","created_at":"2026-07-05T10:44:53.103824+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH","json":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH.json","graph_json":"https://pith.science/api/pith-number/C63I6HX6STBY6BESHFP64WDARH/graph.json","events_json":"https://pith.science/api/pith-number/C63I6HX6STBY6BESHFP64WDARH/events.json","paper":"https://pith.science/paper/C63I6HX6"},"agent_actions":{"view_html":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH","download_json":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH.json","view_paper":"https://pith.science/paper/C63I6HX6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.04024&json=true","fetch_graph":"https://pith.science/api/pith-number/C63I6HX6STBY6BESHFP64WDARH/graph.json","fetch_events":"https://pith.science/api/pith-number/C63I6HX6STBY6BESHFP64WDARH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH/action/storage_attestation","attest_author":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH/action/author_attestation","sign_citation":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH/action/citation_signature","submit_replication":"https://pith.science/pith/C63I6HX6STBY6BESHFP64WDARH/action/replication_record"}},"created_at":"2026-07-05T10:44:53.103824+00:00","updated_at":"2026-07-05T10:44:53.103824+00:00"}