{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KID56CIHUGNMJSSB4SKPI7XGGE","short_pith_number":"pith:KID56CIH","schema_version":"1.0","canonical_sha256":"5207df0907a19ac4ca41e494f47ee631117df4bbb380da26aa0f1afad50e302f","source":{"kind":"arxiv","id":"2502.20339","version":1},"attestation_state":"computed","paper":{"title":"Thinking Slow, Fast: Scaling Inference Compute with Distilled Reasoners","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Albert Gu, Aviv Bick, Daniele Paliotta, Fran\\c{c}ois Fleuret, Junxiong Wang, J. Zico Kolter, Kevin Y. Li, Matteo Pagliardini, Tri Dao","submitted_at":"2025-02-27T18:08:16Z","abstract_excerpt":"Recent advancements have demonstrated that the performance of large language models (LLMs) can be significantly enhanced by scaling computational resources at test time. A common strategy involves generating multiple Chain-of-Thought (CoT) trajectories and aggregating their outputs through various selection mechanisms. This raises a fundamental question: can models with lower complexity leverage their superior generation throughput to outperform similarly sized Transformers for a fixed computational budget? To address this question and overcome the lack of strong subquadratic reasoners, we dis"},"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":"2502.20339","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-27T18:08:16Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e4131a39d115b19c6ee477f3b9ab91387be06d67207ad29de2f505c0e72d8f17","abstract_canon_sha256":"684a28b76471cd7d8bc2342be5acc075011f35e5ae92dc6b65292175a3ef3bed"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:21:13.201579Z","signature_b64":"u7xyW3Bf0AeRAhASbIw6rSgdFjH/8XIelHeucU6P+8SUgdXPCuCTjdcdWdgVFZKQHyS5TAlO8xhiHruZLb2fAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5207df0907a19ac4ca41e494f47ee631117df4bbb380da26aa0f1afad50e302f","last_reissued_at":"2026-07-05T10:21:13.201048Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:21:13.201048Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Thinking Slow, Fast: Scaling Inference Compute with Distilled Reasoners","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Albert Gu, Aviv Bick, Daniele Paliotta, Fran\\c{c}ois Fleuret, Junxiong Wang, J. Zico Kolter, Kevin Y. Li, Matteo Pagliardini, Tri Dao","submitted_at":"2025-02-27T18:08:16Z","abstract_excerpt":"Recent advancements have demonstrated that the performance of large language models (LLMs) can be significantly enhanced by scaling computational resources at test time. A common strategy involves generating multiple Chain-of-Thought (CoT) trajectories and aggregating their outputs through various selection mechanisms. This raises a fundamental question: can models with lower complexity leverage their superior generation throughput to outperform similarly sized Transformers for a fixed computational budget? To address this question and overcome the lack of strong subquadratic reasoners, we dis"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.20339","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/2502.20339/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":"2502.20339","created_at":"2026-07-05T10:21:13.201114+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.20339v1","created_at":"2026-07-05T10:21:13.201114+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.20339","created_at":"2026-07-05T10:21:13.201114+00:00"},{"alias_kind":"pith_short_12","alias_value":"KID56CIHUGNM","created_at":"2026-07-05T10:21:13.201114+00:00"},{"alias_kind":"pith_short_16","alias_value":"KID56CIHUGNMJSSB","created_at":"2026-07-05T10:21:13.201114+00:00"},{"alias_kind":"pith_short_8","alias_value":"KID56CIH","created_at":"2026-07-05T10:21:13.201114+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.30562","citing_title":"Morphing into Hybrid Attention Models","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07604","citing_title":"Contribution Weights: A Geometrical Analysis of Self-Attention Transformers","ref_index":106,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE","json":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE.json","graph_json":"https://pith.science/api/pith-number/KID56CIHUGNMJSSB4SKPI7XGGE/graph.json","events_json":"https://pith.science/api/pith-number/KID56CIHUGNMJSSB4SKPI7XGGE/events.json","paper":"https://pith.science/paper/KID56CIH"},"agent_actions":{"view_html":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE","download_json":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE.json","view_paper":"https://pith.science/paper/KID56CIH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.20339&json=true","fetch_graph":"https://pith.science/api/pith-number/KID56CIHUGNMJSSB4SKPI7XGGE/graph.json","fetch_events":"https://pith.science/api/pith-number/KID56CIHUGNMJSSB4SKPI7XGGE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE/action/storage_attestation","attest_author":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE/action/author_attestation","sign_citation":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE/action/citation_signature","submit_replication":"https://pith.science/pith/KID56CIHUGNMJSSB4SKPI7XGGE/action/replication_record"}},"created_at":"2026-07-05T10:21:13.201114+00:00","updated_at":"2026-07-05T10:21:13.201114+00:00"}