{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:UWWP3LBNPG5O72O22EEIA554HZ","short_pith_number":"pith:UWWP3LBN","schema_version":"1.0","canonical_sha256":"a5acfdac2d79baefe9dad1088077bc3e5e474a56a5db1f56511506329fc0fc2d","source":{"kind":"arxiv","id":"2506.18237","version":1},"attestation_state":"computed","paper":{"title":"AdapThink: Adaptive Thinking Preferences for Reasoning Language Model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Jie Song, Mingyang Sun, Wei Wang, Wenyue Xu, Wotao Yin, Xu Wan","submitted_at":"2025-06-23T02:06:04Z","abstract_excerpt":"Reinforcement Learning (RL)-based post-training has significantly advanced the complex reasoning capabilities of language models, fostering sophisticated self-reflection processes. However, this ``slow thinking'' paradigm presents a critical challenge to reasoning efficiency: models may expend excessive computation on simple questions and shift reasoning prematurely for complex ones. Previous mechanisms typically rely on static length budgets or predefined rules, lacking the adaptability for varying question complexities and models' evolving capabilities. To this end, we propose AdapThink, an "},"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":"2506.18237","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-23T02:06:04Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"08f653370109166de11a1580b87e7e465b1f491e56c75e2240a9d55a88c2f775","abstract_canon_sha256":"455566260f884460fa491ab6fa7593c94ea2cf1f41858463314fb54fc322852e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:25:31.258592Z","signature_b64":"kpnbs4vENvqSeUwVyBOe7VL1+AxFhiwq620f7JVd2JWKSbv1pckdwgIYzp7V4zLkBpbKSjFSonysc0TY9oi8Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a5acfdac2d79baefe9dad1088077bc3e5e474a56a5db1f56511506329fc0fc2d","last_reissued_at":"2026-07-05T11:25:31.258153Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:25:31.258153Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AdapThink: Adaptive Thinking Preferences for Reasoning Language Model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Jie Song, Mingyang Sun, Wei Wang, Wenyue Xu, Wotao Yin, Xu Wan","submitted_at":"2025-06-23T02:06:04Z","abstract_excerpt":"Reinforcement Learning (RL)-based post-training has significantly advanced the complex reasoning capabilities of language models, fostering sophisticated self-reflection processes. However, this ``slow thinking'' paradigm presents a critical challenge to reasoning efficiency: models may expend excessive computation on simple questions and shift reasoning prematurely for complex ones. Previous mechanisms typically rely on static length budgets or predefined rules, lacking the adaptability for varying question complexities and models' evolving capabilities. To this end, we propose AdapThink, an "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.18237","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/2506.18237/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":"2506.18237","created_at":"2026-07-05T11:25:31.258202+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.18237v1","created_at":"2026-07-05T11:25:31.258202+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.18237","created_at":"2026-07-05T11:25:31.258202+00:00"},{"alias_kind":"pith_short_12","alias_value":"UWWP3LBNPG5O","created_at":"2026-07-05T11:25:31.258202+00:00"},{"alias_kind":"pith_short_16","alias_value":"UWWP3LBNPG5O72O2","created_at":"2026-07-05T11:25:31.258202+00:00"},{"alias_kind":"pith_short_8","alias_value":"UWWP3LBN","created_at":"2026-07-05T11:25:31.258202+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05861","citing_title":"Mitigating Factual Hallucination in Large Reasoning Models via Mixed-Mode Advantage Regularization","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2606.18089","citing_title":"From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning","ref_index":119,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10195","citing_title":"Breaking the Reward Barrier: Accelerating Tree-of-Thought Reasoning via Speculative Exploration","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10195","citing_title":"Breaking the Reward Barrier: Accelerating Tree-of-Thought Reasoning via Speculative Exploration","ref_index":53,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ","json":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ.json","graph_json":"https://pith.science/api/pith-number/UWWP3LBNPG5O72O22EEIA554HZ/graph.json","events_json":"https://pith.science/api/pith-number/UWWP3LBNPG5O72O22EEIA554HZ/events.json","paper":"https://pith.science/paper/UWWP3LBN"},"agent_actions":{"view_html":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ","download_json":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ.json","view_paper":"https://pith.science/paper/UWWP3LBN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.18237&json=true","fetch_graph":"https://pith.science/api/pith-number/UWWP3LBNPG5O72O22EEIA554HZ/graph.json","fetch_events":"https://pith.science/api/pith-number/UWWP3LBNPG5O72O22EEIA554HZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ/action/storage_attestation","attest_author":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ/action/author_attestation","sign_citation":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ/action/citation_signature","submit_replication":"https://pith.science/pith/UWWP3LBNPG5O72O22EEIA554HZ/action/replication_record"}},"created_at":"2026-07-05T11:25:31.258202+00:00","updated_at":"2026-07-05T11:25:31.258202+00:00"}