{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:AXLK4TV7ZXMYMC6QVDXDMHMHYL","short_pith_number":"pith:AXLK4TV7","schema_version":"1.0","canonical_sha256":"05d6ae4ebfcdd9860bd0a8ee361d87c2c67e7503a7ff350d5e7fc942c02fd697","source":{"kind":"arxiv","id":"2501.05453","version":1},"attestation_state":"computed","paper":{"title":"An Empirical Study of Autoregressive Pre-training from Videos","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Christoph Feichtenhofer, Ilija Radosavovic, Jathushan Rajasegaran, Jitendra Malik, Rahul Ravishankar, Yossi Gandelsman","submitted_at":"2025-01-09T18:59:58Z","abstract_excerpt":"We empirically study autoregressive pre-training from videos. To perform our study, we construct a series of autoregressive video models, called Toto. We treat videos as sequences of visual tokens and train transformer models to autoregressively predict future tokens. Our models are pre-trained on a diverse dataset of videos and images comprising over 1 trillion visual tokens. We explore different architectural, training, and inference design choices. We evaluate the learned visual representations on a range of downstream tasks including image recognition, video classification, object tracking"},"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":"2501.05453","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2025-01-09T18:59:58Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bfd7d3c21260b5f0783aa90e639a849aed3e5f6737b23a8c0f5e3e726553d2ac","abstract_canon_sha256":"8a627d0c06f34f296724eabec935eed6a9ee94c5a67d5ed4f593716ad45f1cbf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:59:10.118980Z","signature_b64":"zhBKLkod5ldRJuWVNB5yhiOBfYxae9vGG/Jtc7Pun9o96js32/7CCNjXHjM252BmVyfVlvyabfvreXE/xpELBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"05d6ae4ebfcdd9860bd0a8ee361d87c2c67e7503a7ff350d5e7fc942c02fd697","last_reissued_at":"2026-07-05T09:59:10.118499Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:59:10.118499Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An Empirical Study of Autoregressive Pre-training from Videos","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Christoph Feichtenhofer, Ilija Radosavovic, Jathushan Rajasegaran, Jitendra Malik, Rahul Ravishankar, Yossi Gandelsman","submitted_at":"2025-01-09T18:59:58Z","abstract_excerpt":"We empirically study autoregressive pre-training from videos. To perform our study, we construct a series of autoregressive video models, called Toto. We treat videos as sequences of visual tokens and train transformer models to autoregressively predict future tokens. Our models are pre-trained on a diverse dataset of videos and images comprising over 1 trillion visual tokens. We explore different architectural, training, and inference design choices. We evaluate the learned visual representations on a range of downstream tasks including image recognition, video classification, object tracking"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.05453","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/2501.05453/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":"2501.05453","created_at":"2026-07-05T09:59:10.118560+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.05453v1","created_at":"2026-07-05T09:59:10.118560+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.05453","created_at":"2026-07-05T09:59:10.118560+00:00"},{"alias_kind":"pith_short_12","alias_value":"AXLK4TV7ZXMY","created_at":"2026-07-05T09:59:10.118560+00:00"},{"alias_kind":"pith_short_16","alias_value":"AXLK4TV7ZXMYMC6Q","created_at":"2026-07-05T09:59:10.118560+00:00"},{"alias_kind":"pith_short_8","alias_value":"AXLK4TV7","created_at":"2026-07-05T09:59:10.118560+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.23033","citing_title":"Uncovering the Latent Potential of Deep Intermediate Representations","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2507.13942","citing_title":"Frozen Forecasting: A Unified Evaluation","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2504.13181","citing_title":"Perception Encoder: The best visual embeddings are not at the output of the network","ref_index":107,"is_internal_anchor":false},{"citing_arxiv_id":"2506.01844","citing_title":"SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2506.09985","citing_title":"V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning","ref_index":45,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL","json":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL.json","graph_json":"https://pith.science/api/pith-number/AXLK4TV7ZXMYMC6QVDXDMHMHYL/graph.json","events_json":"https://pith.science/api/pith-number/AXLK4TV7ZXMYMC6QVDXDMHMHYL/events.json","paper":"https://pith.science/paper/AXLK4TV7"},"agent_actions":{"view_html":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL","download_json":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL.json","view_paper":"https://pith.science/paper/AXLK4TV7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.05453&json=true","fetch_graph":"https://pith.science/api/pith-number/AXLK4TV7ZXMYMC6QVDXDMHMHYL/graph.json","fetch_events":"https://pith.science/api/pith-number/AXLK4TV7ZXMYMC6QVDXDMHMHYL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL/action/storage_attestation","attest_author":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL/action/author_attestation","sign_citation":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL/action/citation_signature","submit_replication":"https://pith.science/pith/AXLK4TV7ZXMYMC6QVDXDMHMHYL/action/replication_record"}},"created_at":"2026-07-05T09:59:10.118560+00:00","updated_at":"2026-07-05T09:59:10.118560+00:00"}