{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:PMI7K22QU33FM7PXDI2HCLPOYX","short_pith_number":"pith:PMI7K22Q","schema_version":"1.0","canonical_sha256":"7b11f56b50a6f6567df71a34712deec5df5dfeba3acb5013211b8b180ccd96c0","source":{"kind":"arxiv","id":"2505.17952","version":1},"attestation_state":"computed","paper":{"title":"Beyond Distillation: Pushing the Limits of Medical LLM Reasoning with Minimalist Rule-Based RL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Che Liu, Daniel Rueckert, Fangzhen Lin, Haozhe Wang, Jiazhen Pan, Rossella Arcucci, Wenjia Bai, Yong Dai, Zhongwei Wan","submitted_at":"2025-05-23T14:27:37Z","abstract_excerpt":"Improving performance on complex tasks and enabling interpretable decision making in large language models (LLMs), especially for clinical applications, requires effective reasoning. Yet this remains challenging without supervised fine-tuning (SFT) on costly chain-of-thought (CoT) data distilled from closed-source models (e.g., GPT-4o). In this work, we present AlphaMed, the first medical LLM to show that reasoning capability can emerge purely through reinforcement learning (RL), using minimalist rule-based rewards on public multiple-choice QA datasets, without relying on SFT or distilled CoT "},"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":"2505.17952","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-23T14:27:37Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"cec6c24f240bc8a296343df551aed83762bd4d641e6a2613477d1740bd26f586","abstract_canon_sha256":"ba31ec7df7d1ecec7a5007b8cb26001250f7e5c0891d2ad620d55c96b53f4cdb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:08:34.231873Z","signature_b64":"lAdviW73N7vff6WKybfFiKY1A9518ifh90RefnhvaGYmJv/rsx3SLC8Gtvq1HeNlAUlx6zY2fpicfl5XUEYUBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b11f56b50a6f6567df71a34712deec5df5dfeba3acb5013211b8b180ccd96c0","last_reissued_at":"2026-07-05T11:08:34.231420Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:08:34.231420Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond Distillation: Pushing the Limits of Medical LLM Reasoning with Minimalist Rule-Based RL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Che Liu, Daniel Rueckert, Fangzhen Lin, Haozhe Wang, Jiazhen Pan, Rossella Arcucci, Wenjia Bai, Yong Dai, Zhongwei Wan","submitted_at":"2025-05-23T14:27:37Z","abstract_excerpt":"Improving performance on complex tasks and enabling interpretable decision making in large language models (LLMs), especially for clinical applications, requires effective reasoning. Yet this remains challenging without supervised fine-tuning (SFT) on costly chain-of-thought (CoT) data distilled from closed-source models (e.g., GPT-4o). In this work, we present AlphaMed, the first medical LLM to show that reasoning capability can emerge purely through reinforcement learning (RL), using minimalist rule-based rewards on public multiple-choice QA datasets, without relying on SFT or distilled CoT "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.17952","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/2505.17952/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":"2505.17952","created_at":"2026-07-05T11:08:34.231478+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.17952v1","created_at":"2026-07-05T11:08:34.231478+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.17952","created_at":"2026-07-05T11:08:34.231478+00:00"},{"alias_kind":"pith_short_12","alias_value":"PMI7K22QU33F","created_at":"2026-07-05T11:08:34.231478+00:00"},{"alias_kind":"pith_short_16","alias_value":"PMI7K22QU33FM7PX","created_at":"2026-07-05T11:08:34.231478+00:00"},{"alias_kind":"pith_short_8","alias_value":"PMI7K22Q","created_at":"2026-07-05T11:08:34.231478+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.11740","citing_title":"UniReason-Med: A Shared Grounded Reasoning Interface for 2D-to-3D Transfer in Medical VQA","ref_index":186,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08559","citing_title":"Medical Reasoning with Large Language Models: A Survey and MR-Bench","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09584","citing_title":"CLR-voyance: Reinforcing Open-Ended Reasoning for Inpatient Clinical Decision Support with Outcome-Aware Rubrics","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14475","citing_title":"Evo-MedAgent: Beyond One-Shot Diagnosis with Agents That Remember, Reflect, and Improve","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX","json":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX.json","graph_json":"https://pith.science/api/pith-number/PMI7K22QU33FM7PXDI2HCLPOYX/graph.json","events_json":"https://pith.science/api/pith-number/PMI7K22QU33FM7PXDI2HCLPOYX/events.json","paper":"https://pith.science/paper/PMI7K22Q"},"agent_actions":{"view_html":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX","download_json":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX.json","view_paper":"https://pith.science/paper/PMI7K22Q","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.17952&json=true","fetch_graph":"https://pith.science/api/pith-number/PMI7K22QU33FM7PXDI2HCLPOYX/graph.json","fetch_events":"https://pith.science/api/pith-number/PMI7K22QU33FM7PXDI2HCLPOYX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX/action/storage_attestation","attest_author":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX/action/author_attestation","sign_citation":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX/action/citation_signature","submit_replication":"https://pith.science/pith/PMI7K22QU33FM7PXDI2HCLPOYX/action/replication_record"}},"created_at":"2026-07-05T11:08:34.231478+00:00","updated_at":"2026-07-05T11:08:34.231478+00:00"}