{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:UNGBOKWOQ3MXD4OPCLD5CWG34M","short_pith_number":"pith:UNGBOKWO","schema_version":"1.0","canonical_sha256":"a34c172ace86d971f1cf12c7d158dbe31129c9859bc26baf6fbaf3e044d52c76","source":{"kind":"arxiv","id":"2406.02547","version":1},"attestation_state":"computed","paper":{"title":"Leveraging Visual Tokens for Extended Text Contexts in Multi-Modal Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Alex Jinpeng Wang, Lijuan Wang, Linjie Li, Mike Zheng Shou, Min Li, Yiqi Lin","submitted_at":"2024-06-04T17:59:25Z","abstract_excerpt":"Training models with longer in-context lengths is a significant challenge for multimodal model due to substantial GPU memory and computational costs. This exploratory study does not present state-of-the-art models; rather, it introduces an innovative method designed to increase in-context text length in multi-modality large language models (MLLMs) efficiently. We present Visualized In-Context Text Processing (VisInContext), which processes long in-context text using visual tokens. This technique significantly reduces GPU memory usage and floating point operations (FLOPs) for both training and "},"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":"2406.02547","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-06-04T17:59:25Z","cross_cats_sorted":[],"title_canon_sha256":"fe959ceb73493d5af316f97b2613524ed70a220af71c05eee4e154105ae03eb0","abstract_canon_sha256":"d37cf99edec8b946047b5ab05b8e8f238577e5d9ad27acd3faf3be7c15d6a249"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:27:21.113234Z","signature_b64":"nAsnfhptNn0+SDJ/a8T6dKe+GlLMbNa1pcl8s6C3w1brTCX8cQQvnGtAMtEnqLaONITKqHCahnixeZ9IikJmCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a34c172ace86d971f1cf12c7d158dbe31129c9859bc26baf6fbaf3e044d52c76","last_reissued_at":"2026-07-05T08:27:21.112828Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:27:21.112828Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Leveraging Visual Tokens for Extended Text Contexts in Multi-Modal Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Alex Jinpeng Wang, Lijuan Wang, Linjie Li, Mike Zheng Shou, Min Li, Yiqi Lin","submitted_at":"2024-06-04T17:59:25Z","abstract_excerpt":"Training models with longer in-context lengths is a significant challenge for multimodal model due to substantial GPU memory and computational costs. This exploratory study does not present state-of-the-art models; rather, it introduces an innovative method designed to increase in-context text length in multi-modality large language models (MLLMs) efficiently. We present Visualized In-Context Text Processing (VisInContext), which processes long in-context text using visual tokens. This technique significantly reduces GPU memory usage and floating point operations (FLOPs) for both training and "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.02547","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/2406.02547/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":"2406.02547","created_at":"2026-07-05T08:27:21.112884+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.02547v1","created_at":"2026-07-05T08:27:21.112884+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.02547","created_at":"2026-07-05T08:27:21.112884+00:00"},{"alias_kind":"pith_short_12","alias_value":"UNGBOKWOQ3MX","created_at":"2026-07-05T08:27:21.112884+00:00"},{"alias_kind":"pith_short_16","alias_value":"UNGBOKWOQ3MXD4OP","created_at":"2026-07-05T08:27:21.112884+00:00"},{"alias_kind":"pith_short_8","alias_value":"UNGBOKWO","created_at":"2026-07-05T08:27:21.112884+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.18150","citing_title":"Improving MLLM Training Efficiency via Stage-Aware Sparsity","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M","json":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M.json","graph_json":"https://pith.science/api/pith-number/UNGBOKWOQ3MXD4OPCLD5CWG34M/graph.json","events_json":"https://pith.science/api/pith-number/UNGBOKWOQ3MXD4OPCLD5CWG34M/events.json","paper":"https://pith.science/paper/UNGBOKWO"},"agent_actions":{"view_html":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M","download_json":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M.json","view_paper":"https://pith.science/paper/UNGBOKWO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.02547&json=true","fetch_graph":"https://pith.science/api/pith-number/UNGBOKWOQ3MXD4OPCLD5CWG34M/graph.json","fetch_events":"https://pith.science/api/pith-number/UNGBOKWOQ3MXD4OPCLD5CWG34M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M/action/storage_attestation","attest_author":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M/action/author_attestation","sign_citation":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M/action/citation_signature","submit_replication":"https://pith.science/pith/UNGBOKWOQ3MXD4OPCLD5CWG34M/action/replication_record"}},"created_at":"2026-07-05T08:27:21.112884+00:00","updated_at":"2026-07-05T08:27:21.112884+00:00"}