{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MVRYENMOKNLLGFW2D76LPSTLID","short_pith_number":"pith:MVRYENMO","schema_version":"1.0","canonical_sha256":"656382358e5356b316da1ffcb7ca6b40fd147c8bedb89b3c3ecbe6dad558763b","source":{"kind":"arxiv","id":"2410.09918","version":3},"attestation_state":"computed","paper":{"title":"Dualformer: Controllable Fast and Slow Thinking by Learning with Randomized Reasoning Traces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.LO"],"primary_cat":"cs.AI","authors_text":"DiJia Su, Michael Rabbat, Qinqing Zheng, Sainbayar Sukhbaatar, Yuandong Tian","submitted_at":"2024-10-13T16:53:02Z","abstract_excerpt":"In cognition theory, human thinking is governed by two systems: the fast and intuitive System 1 and the slower but more deliberative System 2. Analogously, Large Language Models (LLMs) can operate in two reasoning modes: outputting only the solutions (\\emph{fast mode}) or both the reasoning chain and the final solution (\\emph{slow mode}). We present \\dualformer, a single Transformer model that seamlessly integrates both the fast and slow reasoning modes by training on randomized reasoning traces, where different parts of the traces are strategically dropped during training. At inference time, "},"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":"2410.09918","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-10-13T16:53:02Z","cross_cats_sorted":["cs.LG","cs.LO"],"title_canon_sha256":"230d0a1f5a3d4ba607598f9d0329abdba1de69fe775bca4ae1027d1bb669719d","abstract_canon_sha256":"665862ecf7cae6cb49b1048325537d204d2cbf0212f9bad71a28f01c0ef5dbae"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:35:10.837598Z","signature_b64":"5O+oFP1qdJfiuYAxMGRFz34Wes0u0I1GJ0ntitJ+uj81h164FMTqFeWp/+u7IE8z+QblCbufkbR+peysU20DAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"656382358e5356b316da1ffcb7ca6b40fd147c8bedb89b3c3ecbe6dad558763b","last_reissued_at":"2026-07-05T11:35:10.837083Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:35:10.837083Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dualformer: Controllable Fast and Slow Thinking by Learning with Randomized Reasoning Traces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.LO"],"primary_cat":"cs.AI","authors_text":"DiJia Su, Michael Rabbat, Qinqing Zheng, Sainbayar Sukhbaatar, Yuandong Tian","submitted_at":"2024-10-13T16:53:02Z","abstract_excerpt":"In cognition theory, human thinking is governed by two systems: the fast and intuitive System 1 and the slower but more deliberative System 2. Analogously, Large Language Models (LLMs) can operate in two reasoning modes: outputting only the solutions (\\emph{fast mode}) or both the reasoning chain and the final solution (\\emph{slow mode}). We present \\dualformer, a single Transformer model that seamlessly integrates both the fast and slow reasoning modes by training on randomized reasoning traces, where different parts of the traces are strategically dropped during training. At inference time, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.09918","kind":"arxiv","version":3},"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/2410.09918/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":"2410.09918","created_at":"2026-07-05T11:35:10.837140+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.09918v3","created_at":"2026-07-05T11:35:10.837140+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.09918","created_at":"2026-07-05T11:35:10.837140+00:00"},{"alias_kind":"pith_short_12","alias_value":"MVRYENMOKNLL","created_at":"2026-07-05T11:35:10.837140+00:00"},{"alias_kind":"pith_short_16","alias_value":"MVRYENMOKNLLGFW2","created_at":"2026-07-05T11:35:10.837140+00:00"},{"alias_kind":"pith_short_8","alias_value":"MVRYENMO","created_at":"2026-07-05T11:35:10.837140+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07492","citing_title":"Search, Fail, Recover: A Training Framework for Correction-Aware Reasoning","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2501.19201","citing_title":"Efficient Reasoning with Hidden Thinking","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2501.09732","citing_title":"Inference-Time Scaling for Diffusion Models beyond Scaling Denoising Steps","ref_index":74,"is_internal_anchor":false},{"citing_arxiv_id":"2602.04476","citing_title":"Vision-aligned Latent Reasoning for Multi-modal Large Language Model","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08221","citing_title":"NoisyCoconut: Counterfactual Consensus via Latent Space Reasoning","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2412.06769","citing_title":"Training Large Language Models to Reason in a Continuous Latent Space","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21027","citing_title":"HypEHR: Hyperbolic Modeling of Electronic Health Records for Efficient Question Answering","ref_index":82,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID","json":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID.json","graph_json":"https://pith.science/api/pith-number/MVRYENMOKNLLGFW2D76LPSTLID/graph.json","events_json":"https://pith.science/api/pith-number/MVRYENMOKNLLGFW2D76LPSTLID/events.json","paper":"https://pith.science/paper/MVRYENMO"},"agent_actions":{"view_html":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID","download_json":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID.json","view_paper":"https://pith.science/paper/MVRYENMO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.09918&json=true","fetch_graph":"https://pith.science/api/pith-number/MVRYENMOKNLLGFW2D76LPSTLID/graph.json","fetch_events":"https://pith.science/api/pith-number/MVRYENMOKNLLGFW2D76LPSTLID/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID/action/storage_attestation","attest_author":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID/action/author_attestation","sign_citation":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID/action/citation_signature","submit_replication":"https://pith.science/pith/MVRYENMOKNLLGFW2D76LPSTLID/action/replication_record"}},"created_at":"2026-07-05T11:35:10.837140+00:00","updated_at":"2026-07-05T11:35:10.837140+00:00"}