{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MPW2WC7J3STYDB3XTT3MGKSZCE","short_pith_number":"pith:MPW2WC7J","schema_version":"1.0","canonical_sha256":"63edab0be9dca78187779cf6c32a5911263d83e2a74b87b252ed6630569cb16c","source":{"kind":"arxiv","id":"2406.00977","version":2},"attestation_state":"computed","paper":{"title":"Dragonfly: Multi-Resolution Zoom-In Encoding Enhances Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Ben Athiwaratkun, Ian Covert, James Zou, Kezhen Chen, Rahul Chalamala, Rahul Thapa, Shuaiwen Leon Song","submitted_at":"2024-06-03T04:17:12Z","abstract_excerpt":"Recent advances in vision-language models (VLMs) have demonstrated the advantages of processing images at higher resolutions and utilizing multi-crop features to preserve native resolution details. However, despite these improvements, existing vision transformers (ViTs) still struggle to capture fine-grained details from less prominent objects, charts, and embedded text, limiting their effectiveness in certain tasks. In this paper, we extend recent high-resolution and multi-crop techniques by not only preserving the native resolution, but zooming in beyond it and extracting features from a lar"},"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":"2406.00977","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-06-03T04:17:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8cdb2a573f3c14003610cd67828f052e00ec263624971f4c0291a31e2accd1fc","abstract_canon_sha256":"9a169005ab4cc98a77bd1cc160f2609cf698cccbfb5fa0547ce86ba7f0a9cebf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:20:23.591007Z","signature_b64":"q05Pq5EDXLSw7gDlXjtM8UWUJn1VpybiHvaouHoroNOGS7P5dHrmzLp0oBqkGUKQBIaz3HDT//Nuk9JzP+LSAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"63edab0be9dca78187779cf6c32a5911263d83e2a74b87b252ed6630569cb16c","last_reissued_at":"2026-07-05T09:20:23.590468Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:20:23.590468Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dragonfly: Multi-Resolution Zoom-In Encoding Enhances Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Ben Athiwaratkun, Ian Covert, James Zou, Kezhen Chen, Rahul Chalamala, Rahul Thapa, Shuaiwen Leon Song","submitted_at":"2024-06-03T04:17:12Z","abstract_excerpt":"Recent advances in vision-language models (VLMs) have demonstrated the advantages of processing images at higher resolutions and utilizing multi-crop features to preserve native resolution details. However, despite these improvements, existing vision transformers (ViTs) still struggle to capture fine-grained details from less prominent objects, charts, and embedded text, limiting their effectiveness in certain tasks. In this paper, we extend recent high-resolution and multi-crop techniques by not only preserving the native resolution, but zooming in beyond it and extracting features from a lar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.00977","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/2406.00977/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":"2406.00977","created_at":"2026-07-05T09:20:23.590530+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.00977v2","created_at":"2026-07-05T09:20:23.590530+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.00977","created_at":"2026-07-05T09:20:23.590530+00:00"},{"alias_kind":"pith_short_12","alias_value":"MPW2WC7J3STY","created_at":"2026-07-05T09:20:23.590530+00:00"},{"alias_kind":"pith_short_16","alias_value":"MPW2WC7J3STYDB3X","created_at":"2026-07-05T09:20:23.590530+00:00"},{"alias_kind":"pith_short_8","alias_value":"MPW2WC7J","created_at":"2026-07-05T09:20:23.590530+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.22098","citing_title":"TextTeacher: What Can Language Teach About Images?","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2409.17146","citing_title":"Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models","ref_index":105,"is_internal_anchor":false},{"citing_arxiv_id":"2504.07615","citing_title":"VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07019","citing_title":"LensVLM: Selective Context Expansion for Compressed Visual Representation of Text","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE","json":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE.json","graph_json":"https://pith.science/api/pith-number/MPW2WC7J3STYDB3XTT3MGKSZCE/graph.json","events_json":"https://pith.science/api/pith-number/MPW2WC7J3STYDB3XTT3MGKSZCE/events.json","paper":"https://pith.science/paper/MPW2WC7J"},"agent_actions":{"view_html":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE","download_json":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE.json","view_paper":"https://pith.science/paper/MPW2WC7J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.00977&json=true","fetch_graph":"https://pith.science/api/pith-number/MPW2WC7J3STYDB3XTT3MGKSZCE/graph.json","fetch_events":"https://pith.science/api/pith-number/MPW2WC7J3STYDB3XTT3MGKSZCE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE/action/storage_attestation","attest_author":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE/action/author_attestation","sign_citation":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE/action/citation_signature","submit_replication":"https://pith.science/pith/MPW2WC7J3STYDB3XTT3MGKSZCE/action/replication_record"}},"created_at":"2026-07-05T09:20:23.590530+00:00","updated_at":"2026-07-05T09:20:23.590530+00:00"}