{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:5PAEBXSGZ5Z3TKXOYKZ23KAFQY","short_pith_number":"pith:5PAEBXSG","schema_version":"1.0","canonical_sha256":"ebc040de46cf73b9aaeec2b3ada80586229b5ce0f3a56999156863e89ded0b32","source":{"kind":"arxiv","id":"2607.03624","version":1},"attestation_state":"computed","paper":{"title":"RADIO1D: Elastic Representations for Condensed Vision Modeling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Andrew Tao, Bryan Catanzaro, Collin McCarthy, Eugene Khvedchenya, Greg Heinrich, Jan Kautz, Mike Ranzinger, Natan Bagrov, Pavlo Molchanov","submitted_at":"2026-07-03T22:58:54Z","abstract_excerpt":"This paper challenges the assumption that vision-language models (VLMs) require fixed patch-based 2D vision features. Analyzing fine-tuned vision encoders, we find that representations become increasingly abstract and less spatially coherent during VLM training. Notably, models trained with image-text alignment (such as SigLIP2) develop a small number of specialized tokens that effectively summarize global image content. Building on this, we introduce RADIO1D, which compresses images into a compact, variable-length 1D token sequence using multi-teacher knowledge distillation and an autoencoder"},"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":"2607.03624","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-07-03T22:58:54Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"3032f3145cf685013ea58c0da0a494d4fc240d0e6bee04f6d4a5fd1ebb510874","abstract_canon_sha256":"f34d32dcf98ef62f45690e3400648e8b3dcee8b8f01ab09cb0e8485eacb98781"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:17:58.006946Z","signature_b64":"hYbFk7jz3HCOaUKDkZLnuhMidaFJdM+0zow50nSv9rNKT2A6ejwXnKyVZSYYQcnowF9Ak49SD1uDGTySV9lrCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ebc040de46cf73b9aaeec2b3ada80586229b5ce0f3a56999156863e89ded0b32","last_reissued_at":"2026-07-07T02:17:58.005994Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:17:58.005994Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RADIO1D: Elastic Representations for Condensed Vision Modeling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Andrew Tao, Bryan Catanzaro, Collin McCarthy, Eugene Khvedchenya, Greg Heinrich, Jan Kautz, Mike Ranzinger, Natan Bagrov, Pavlo Molchanov","submitted_at":"2026-07-03T22:58:54Z","abstract_excerpt":"This paper challenges the assumption that vision-language models (VLMs) require fixed patch-based 2D vision features. Analyzing fine-tuned vision encoders, we find that representations become increasingly abstract and less spatially coherent during VLM training. Notably, models trained with image-text alignment (such as SigLIP2) develop a small number of specialized tokens that effectively summarize global image content. Building on this, we introduce RADIO1D, which compresses images into a compact, variable-length 1D token sequence using multi-teacher knowledge distillation and an autoencoder"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.03624","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/2607.03624/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":"2607.03624","created_at":"2026-07-07T02:17:58.006107+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.03624v1","created_at":"2026-07-07T02:17:58.006107+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.03624","created_at":"2026-07-07T02:17:58.006107+00:00"},{"alias_kind":"pith_short_12","alias_value":"5PAEBXSGZ5Z3","created_at":"2026-07-07T02:17:58.006107+00:00"},{"alias_kind":"pith_short_16","alias_value":"5PAEBXSGZ5Z3TKXO","created_at":"2026-07-07T02:17:58.006107+00:00"},{"alias_kind":"pith_short_8","alias_value":"5PAEBXSG","created_at":"2026-07-07T02:17:58.006107+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/5PAEBXSGZ5Z3TKXOYKZ23KAFQY","json":"https://pith.science/pith/5PAEBXSGZ5Z3TKXOYKZ23KAFQY.json","graph_json":"https://pith.science/api/pith-number/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/graph.json","events_json":"https://pith.science/api/pith-number/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/events.json","paper":"https://pith.science/paper/5PAEBXSG"},"agent_actions":{"view_html":"https://pith.science/pith/5PAEBXSGZ5Z3TKXOYKZ23KAFQY","download_json":"https://pith.science/pith/5PAEBXSGZ5Z3TKXOYKZ23KAFQY.json","view_paper":"https://pith.science/paper/5PAEBXSG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.03624&json=true","fetch_graph":"https://pith.science/api/pith-number/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/graph.json","fetch_events":"https://pith.science/api/pith-number/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/action/storage_attestation","attest_author":"https://pith.science/pith/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/action/author_attestation","sign_citation":"https://pith.science/pith/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/action/citation_signature","submit_replication":"https://pith.science/pith/5PAEBXSGZ5Z3TKXOYKZ23KAFQY/action/replication_record"}},"created_at":"2026-07-07T02:17:58.006107+00:00","updated_at":"2026-07-07T02:17:58.006107+00:00"}