{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JAQ7W3LZEE7GFQFKG2SO2LY5J3","short_pith_number":"pith:JAQ7W3LZ","schema_version":"1.0","canonical_sha256":"4821fb6d79213e62c0aa36a4ed2f1d4ed6c7c83723fa4e56fe9e604ed1ba6d86","source":{"kind":"arxiv","id":"2505.21178","version":1},"attestation_state":"computed","paper":{"title":"Walk Before You Run! Concise LLM Reasoning via Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Mao Zheng, Mingyang Song","submitted_at":"2025-05-27T13:29:51Z","abstract_excerpt":"As test-time scaling becomes a pivotal research frontier in Large Language Models (LLMs) development, contemporary and advanced post-training methodologies increasingly focus on extending the generation length of long Chain-of-Thought (CoT) responses to enhance reasoning capabilities toward DeepSeek R1-like performance. However, recent studies reveal a persistent overthinking phenomenon in state-of-the-art reasoning models, manifesting as excessive redundancy or repetitive thinking patterns in long CoT responses. To address this issue, in this paper, we propose a simple yet effective two-stage"},"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":"2505.21178","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-27T13:29:51Z","cross_cats_sorted":[],"title_canon_sha256":"614b157bc3ca1156f7769b25d08322740664ccb74a1f250cc53b9a1e832c8524","abstract_canon_sha256":"d4f32cbac89e9927c211c7bbe9c7c399fa8fba1ec680b618cca74a3cd3c6838e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:10:25.059156Z","signature_b64":"24w9DxBeaMhgGfWLxzhLzh8H6lrA9hwoTou1xQY0wlsH9zWEpSQ3zlydDb/2aniSiSOeQlFJ9IjiyfrKI3GADQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4821fb6d79213e62c0aa36a4ed2f1d4ed6c7c83723fa4e56fe9e604ed1ba6d86","last_reissued_at":"2026-07-05T11:10:25.058710Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:10:25.058710Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Walk Before You Run! Concise LLM Reasoning via Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Mao Zheng, Mingyang Song","submitted_at":"2025-05-27T13:29:51Z","abstract_excerpt":"As test-time scaling becomes a pivotal research frontier in Large Language Models (LLMs) development, contemporary and advanced post-training methodologies increasingly focus on extending the generation length of long Chain-of-Thought (CoT) responses to enhance reasoning capabilities toward DeepSeek R1-like performance. However, recent studies reveal a persistent overthinking phenomenon in state-of-the-art reasoning models, manifesting as excessive redundancy or repetitive thinking patterns in long CoT responses. To address this issue, in this paper, we propose a simple yet effective two-stage"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.21178","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/2505.21178/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":"2505.21178","created_at":"2026-07-05T11:10:25.058762+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.21178v1","created_at":"2026-07-05T11:10:25.058762+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.21178","created_at":"2026-07-05T11:10:25.058762+00:00"},{"alias_kind":"pith_short_12","alias_value":"JAQ7W3LZEE7G","created_at":"2026-07-05T11:10:25.058762+00:00"},{"alias_kind":"pith_short_16","alias_value":"JAQ7W3LZEE7GFQFK","created_at":"2026-07-05T11:10:25.058762+00:00"},{"alias_kind":"pith_short_8","alias_value":"JAQ7W3LZ","created_at":"2026-07-05T11:10:25.058762+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22716","citing_title":"Beyond Penalizing Mistakes: Stabilizing Efficiency Training in Large Reasoning Models via Adaptive Correct-Only Rewards","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20295","citing_title":"Token-Operations-Oriented Inference Optimization Techniques for Large Models","ref_index":91,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19358","citing_title":"Taming the Thinker: Conditional Entropy Shaping for Adaptive LLM Reasoning","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2503.16419","citing_title":"Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models","ref_index":160,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06165","citing_title":"Post Reasoning: Improving the Performance of Non-Thinking Models at No Cost","ref_index":251,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01111","citing_title":"When Less is Enough: Efficient Inference via Collaborative Reasoning","ref_index":36,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3","json":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3.json","graph_json":"https://pith.science/api/pith-number/JAQ7W3LZEE7GFQFKG2SO2LY5J3/graph.json","events_json":"https://pith.science/api/pith-number/JAQ7W3LZEE7GFQFKG2SO2LY5J3/events.json","paper":"https://pith.science/paper/JAQ7W3LZ"},"agent_actions":{"view_html":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3","download_json":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3.json","view_paper":"https://pith.science/paper/JAQ7W3LZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.21178&json=true","fetch_graph":"https://pith.science/api/pith-number/JAQ7W3LZEE7GFQFKG2SO2LY5J3/graph.json","fetch_events":"https://pith.science/api/pith-number/JAQ7W3LZEE7GFQFKG2SO2LY5J3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3/action/storage_attestation","attest_author":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3/action/author_attestation","sign_citation":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3/action/citation_signature","submit_replication":"https://pith.science/pith/JAQ7W3LZEE7GFQFKG2SO2LY5J3/action/replication_record"}},"created_at":"2026-07-05T11:10:25.058762+00:00","updated_at":"2026-07-05T11:10:25.058762+00:00"}