{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JXMPVXQJK54DLWUSF7WFHJOCPU","short_pith_number":"pith:JXMPVXQJ","schema_version":"1.0","canonical_sha256":"4dd8fade09577835da922fec53a5c27d21417ef75cbfbd2335e58b6474c32cc9","source":{"kind":"arxiv","id":"2504.00810","version":1},"attestation_state":"computed","paper":{"title":"Z1: Efficient Test-time Scaling with Code","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Arman Cohan, Xiao-Ping Zhang, Yilun Zhao, Yinghao Wu, Zhaojian Yu","submitted_at":"2025-04-01T14:01:50Z","abstract_excerpt":"Large Language Models (LLMs) can achieve enhanced complex problem-solving through test-time computing scaling, yet this often entails longer contexts and numerous reasoning token costs. In this paper, we propose an efficient test-time scaling method that trains LLMs on code-related reasoning trajectories, facilitating their reduction of excess thinking tokens while maintaining performance. First, we create Z1-Code-Reasoning-107K, a curated dataset of simple and complex coding problems paired with their short and long solution trajectories. Second, we present a novel Shifted Thinking Window to "},"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":"2504.00810","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-01T14:01:50Z","cross_cats_sorted":[],"title_canon_sha256":"4b0a9b339a5c68e4c52fb69bfcd7f0638832a411044e08f8b053d8266d6f633e","abstract_canon_sha256":"21431fda3ef7a7ee989ebd86e2a3fba2a609336acf30701a6ced83c4bdead805"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:42:50.760272Z","signature_b64":"kOfmn32JaEs82xSxNzZSNBdF36CKylargfb9KlL4jParCe6FQK+tTL60DYN0Hf0mxZakpqf9NBFB6bMiIybzBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4dd8fade09577835da922fec53a5c27d21417ef75cbfbd2335e58b6474c32cc9","last_reissued_at":"2026-07-05T10:42:50.759764Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:42:50.759764Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Z1: Efficient Test-time Scaling with Code","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Arman Cohan, Xiao-Ping Zhang, Yilun Zhao, Yinghao Wu, Zhaojian Yu","submitted_at":"2025-04-01T14:01:50Z","abstract_excerpt":"Large Language Models (LLMs) can achieve enhanced complex problem-solving through test-time computing scaling, yet this often entails longer contexts and numerous reasoning token costs. In this paper, we propose an efficient test-time scaling method that trains LLMs on code-related reasoning trajectories, facilitating their reduction of excess thinking tokens while maintaining performance. First, we create Z1-Code-Reasoning-107K, a curated dataset of simple and complex coding problems paired with their short and long solution trajectories. Second, we present a novel Shifted Thinking Window to "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.00810","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/2504.00810/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":"2504.00810","created_at":"2026-07-05T10:42:50.759829+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.00810v1","created_at":"2026-07-05T10:42:50.759829+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.00810","created_at":"2026-07-05T10:42:50.759829+00:00"},{"alias_kind":"pith_short_12","alias_value":"JXMPVXQJK54D","created_at":"2026-07-05T10:42:50.759829+00:00"},{"alias_kind":"pith_short_16","alias_value":"JXMPVXQJK54DLWUS","created_at":"2026-07-05T10:42:50.759829+00:00"},{"alias_kind":"pith_short_8","alias_value":"JXMPVXQJ","created_at":"2026-07-05T10:42:50.759829+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.01249","citing_title":"Trust Region On-Policy Distillation","ref_index":295,"is_internal_anchor":false},{"citing_arxiv_id":"2603.27098","citing_title":"Ensemble-Based Uncertainty Estimation for Code Correctness Estimation","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17304","citing_title":"Efficient Test-Time Scaling via Temporal Reasoning Aggregation","ref_index":79,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU","json":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU.json","graph_json":"https://pith.science/api/pith-number/JXMPVXQJK54DLWUSF7WFHJOCPU/graph.json","events_json":"https://pith.science/api/pith-number/JXMPVXQJK54DLWUSF7WFHJOCPU/events.json","paper":"https://pith.science/paper/JXMPVXQJ"},"agent_actions":{"view_html":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU","download_json":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU.json","view_paper":"https://pith.science/paper/JXMPVXQJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.00810&json=true","fetch_graph":"https://pith.science/api/pith-number/JXMPVXQJK54DLWUSF7WFHJOCPU/graph.json","fetch_events":"https://pith.science/api/pith-number/JXMPVXQJK54DLWUSF7WFHJOCPU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU/action/storage_attestation","attest_author":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU/action/author_attestation","sign_citation":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU/action/citation_signature","submit_replication":"https://pith.science/pith/JXMPVXQJK54DLWUSF7WFHJOCPU/action/replication_record"}},"created_at":"2026-07-05T10:42:50.759829+00:00","updated_at":"2026-07-05T10:42:50.759829+00:00"}