{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7QYJMUUVKAQBFOVTGDN7SQIUNW","short_pith_number":"pith:7QYJMUUV","schema_version":"1.0","canonical_sha256":"fc30965295502012bab330dbf941146da79f1ce56227fd76765b249ada15a70f","source":{"kind":"arxiv","id":"2507.01921","version":1},"attestation_state":"computed","paper":{"title":"NaturalThoughts: Selecting and Distilling Reasoning Traces for General Reasoning Tasks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dong Wang, Ilia Kulikov, Jack Lanchantin, Jason Weston, Karthik Padthe, Shang-Wen Li, Thao Nguyen, Weizhe Yuan, Xian Li, Yang Li, Youssef Emad","submitted_at":"2025-07-02T17:30:24Z","abstract_excerpt":"Recent work has shown that distilling reasoning traces from a larger teacher model via supervised finetuning outperforms reinforcement learning with the smaller student model alone (Guo et al. 2025). However, there has not been a systematic study of what kind of reasoning demonstrations from the teacher are most effective in improving the student model's reasoning capabilities. In this work we curate high-quality \"NaturalThoughts\" by selecting reasoning traces from a strong teacher model based on a large pool of questions from NaturalReasoning (Yuan et al. 2025). We first conduct a systematic "},"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":"2507.01921","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-07-02T17:30:24Z","cross_cats_sorted":[],"title_canon_sha256":"ab7b67767a668414949a08463e78fbf7ab20bd04cc5a51a8a256519d759053ad","abstract_canon_sha256":"6acb278fe32398550e963c11ed5fc316f55fa49d2c6ca4d9980ae16ebafad382"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:30:58.271724Z","signature_b64":"SjVHj7T+VL5ZpOCKUSGRAWar799nUQGDv2hs+638OJubZb2wMTNBLhRJVD5R3B9tzl7uoef0mk6mgSrA2HXdAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fc30965295502012bab330dbf941146da79f1ce56227fd76765b249ada15a70f","last_reissued_at":"2026-07-05T11:30:58.271210Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:30:58.271210Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"NaturalThoughts: Selecting and Distilling Reasoning Traces for General Reasoning Tasks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dong Wang, Ilia Kulikov, Jack Lanchantin, Jason Weston, Karthik Padthe, Shang-Wen Li, Thao Nguyen, Weizhe Yuan, Xian Li, Yang Li, Youssef Emad","submitted_at":"2025-07-02T17:30:24Z","abstract_excerpt":"Recent work has shown that distilling reasoning traces from a larger teacher model via supervised finetuning outperforms reinforcement learning with the smaller student model alone (Guo et al. 2025). However, there has not been a systematic study of what kind of reasoning demonstrations from the teacher are most effective in improving the student model's reasoning capabilities. In this work we curate high-quality \"NaturalThoughts\" by selecting reasoning traces from a strong teacher model based on a large pool of questions from NaturalReasoning (Yuan et al. 2025). We first conduct a systematic "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.01921","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/2507.01921/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":"2507.01921","created_at":"2026-07-05T11:30:58.271271+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.01921v1","created_at":"2026-07-05T11:30:58.271271+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.01921","created_at":"2026-07-05T11:30:58.271271+00:00"},{"alias_kind":"pith_short_12","alias_value":"7QYJMUUVKAQB","created_at":"2026-07-05T11:30:58.271271+00:00"},{"alias_kind":"pith_short_16","alias_value":"7QYJMUUVKAQBFOVT","created_at":"2026-07-05T11:30:58.271271+00:00"},{"alias_kind":"pith_short_8","alias_value":"7QYJMUUV","created_at":"2026-07-05T11:30:58.271271+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08268","citing_title":"Different Teachers, Different Capabilities: Sub-1B On-Device Distillation for Structured Text Enrichment","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2605.13165","citing_title":"STOP: Structured On-Policy Pruning of Long-Form Reasoning in Low-Data Regimes","ref_index":77,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02290","citing_title":"Distilling Long-CoT Reasoning through Collaborative Step-wise Multi-Teacher Decoding","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17614","citing_title":"Characterizing Model-Native Skills","ref_index":83,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW","json":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW.json","graph_json":"https://pith.science/api/pith-number/7QYJMUUVKAQBFOVTGDN7SQIUNW/graph.json","events_json":"https://pith.science/api/pith-number/7QYJMUUVKAQBFOVTGDN7SQIUNW/events.json","paper":"https://pith.science/paper/7QYJMUUV"},"agent_actions":{"view_html":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW","download_json":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW.json","view_paper":"https://pith.science/paper/7QYJMUUV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.01921&json=true","fetch_graph":"https://pith.science/api/pith-number/7QYJMUUVKAQBFOVTGDN7SQIUNW/graph.json","fetch_events":"https://pith.science/api/pith-number/7QYJMUUVKAQBFOVTGDN7SQIUNW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW/action/storage_attestation","attest_author":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW/action/author_attestation","sign_citation":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW/action/citation_signature","submit_replication":"https://pith.science/pith/7QYJMUUVKAQBFOVTGDN7SQIUNW/action/replication_record"}},"created_at":"2026-07-05T11:30:58.271271+00:00","updated_at":"2026-07-05T11:30:58.271271+00:00"}