{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:GDHDKFTZTTAW2FLVWFNMMVKL4Y","short_pith_number":"pith:GDHDKFTZ","schema_version":"1.0","canonical_sha256":"30ce3516799cc16d1575b15ac6554be6056d5d1346a731895a8d98470d52f5eb","source":{"kind":"arxiv","id":"2305.12272","version":1},"attestation_state":"computed","paper":{"title":"Autoregressive Modeling with Lookahead Attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Hongyuan Mei, Jason Eisner, Li Du","submitted_at":"2023-05-20T19:29:47Z","abstract_excerpt":"To predict the next token, autoregressive models ordinarily examine the past. Could they also benefit from also examining hypothetical futures? We consider a novel Transformer-based autoregressive architecture that estimates the next-token distribution by extrapolating multiple continuations of the past, according to some proposal distribution, and attending to these extended strings. This architecture draws insights from classical AI systems such as board game players: when making a local decision, a policy may benefit from exploring possible future trajectories and analyzing them. On multipl"},"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":"2305.12272","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-20T19:29:47Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"92d60a684c814b1e548a9eb79cb88bf96d4a031de720fca7f36858e430312855","abstract_canon_sha256":"b01f881c1a271c6bf45c8bc223f374bc3f605b9a122e1c02a925237e4b568223"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:12:08.698607Z","signature_b64":"OgdSz7tXHiWUM/Ky6CLCEO5bn6LVTAowIhOvnGsexkUcsrU6dzfd4yBXdL4NVYXZDaLbPTUwV7Zozu/xoaGGDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"30ce3516799cc16d1575b15ac6554be6056d5d1346a731895a8d98470d52f5eb","last_reissued_at":"2026-07-05T06:12:08.698182Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:12:08.698182Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Autoregressive Modeling with Lookahead Attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Hongyuan Mei, Jason Eisner, Li Du","submitted_at":"2023-05-20T19:29:47Z","abstract_excerpt":"To predict the next token, autoregressive models ordinarily examine the past. Could they also benefit from also examining hypothetical futures? We consider a novel Transformer-based autoregressive architecture that estimates the next-token distribution by extrapolating multiple continuations of the past, according to some proposal distribution, and attending to these extended strings. This architecture draws insights from classical AI systems such as board game players: when making a local decision, a policy may benefit from exploring possible future trajectories and analyzing them. On multipl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.12272","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/2305.12272/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":"2305.12272","created_at":"2026-07-05T06:12:08.698238+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.12272v1","created_at":"2026-07-05T06:12:08.698238+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.12272","created_at":"2026-07-05T06:12:08.698238+00:00"},{"alias_kind":"pith_short_12","alias_value":"GDHDKFTZTTAW","created_at":"2026-07-05T06:12:08.698238+00:00"},{"alias_kind":"pith_short_16","alias_value":"GDHDKFTZTTAW2FLV","created_at":"2026-07-05T06:12:08.698238+00:00"},{"alias_kind":"pith_short_8","alias_value":"GDHDKFTZ","created_at":"2026-07-05T06:12:08.698238+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.13070","citing_title":"Reinforced Context Order Recovery for Adaptive Reasoning and Planning","ref_index":25,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y","json":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y.json","graph_json":"https://pith.science/api/pith-number/GDHDKFTZTTAW2FLVWFNMMVKL4Y/graph.json","events_json":"https://pith.science/api/pith-number/GDHDKFTZTTAW2FLVWFNMMVKL4Y/events.json","paper":"https://pith.science/paper/GDHDKFTZ"},"agent_actions":{"view_html":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y","download_json":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y.json","view_paper":"https://pith.science/paper/GDHDKFTZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.12272&json=true","fetch_graph":"https://pith.science/api/pith-number/GDHDKFTZTTAW2FLVWFNMMVKL4Y/graph.json","fetch_events":"https://pith.science/api/pith-number/GDHDKFTZTTAW2FLVWFNMMVKL4Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y/action/storage_attestation","attest_author":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y/action/author_attestation","sign_citation":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y/action/citation_signature","submit_replication":"https://pith.science/pith/GDHDKFTZTTAW2FLVWFNMMVKL4Y/action/replication_record"}},"created_at":"2026-07-05T06:12:08.698238+00:00","updated_at":"2026-07-05T06:12:08.698238+00:00"}