{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GSQZVRQPA6EV3USSCVSHNUMRJM","short_pith_number":"pith:GSQZVRQP","schema_version":"1.0","canonical_sha256":"34a19ac60f07895dd252156476d1914b34bad88280f082b58c74a05664f629b9","source":{"kind":"arxiv","id":"2509.22193","version":2},"attestation_state":"computed","paper":{"title":"Scale or Reason? A Compute-Equivalent Analysis of Reasoning Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"C\\'eline Hudelot, Hippolyte Gisserot-Boukhlef, Kevin El Haddad, Nicolas Boizard, Pierre Colombo","submitted_at":"2025-09-26T10:53:52Z","abstract_excerpt":"Distilling reasoning traces from strong teacher models has become the standard recipe for building capable small language models. Yet reasoning traces are 5-20$\\times$ longer than standard instruction fine-tuning (IFT) outputs, meaning every practitioner who chooses reasoning distillation implicitly forgoes training a larger IFT model on the same compute budget. Whether this trade-off is worthwhile remains unaddressed. We study it with a controlled experiment: a single teacher generates paired IFT and reasoning outputs for identical prompts by toggling only its reasoning mode, isolating superv"},"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":"2509.22193","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-26T10:53:52Z","cross_cats_sorted":[],"title_canon_sha256":"e5bd2296ef3b6a5e4216a0ad98dfbfd26be46bd4195b85394a7ed874a1ddd714","abstract_canon_sha256":"21740fdd5c209926dd8555ecac327c71f39a3b8dc4ad73a7f6456a0da7fcbcfb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-25T01:17:46.391415Z","signature_b64":"3YgjrZDZ2ZaGKcBMd7M+Q+IR7JKYyHb1sQ1E75Kkf6NNW2D6B1/YmtlByUuVdSOdIsCL1A2VGcTqJUhHBqeJBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"34a19ac60f07895dd252156476d1914b34bad88280f082b58c74a05664f629b9","last_reissued_at":"2026-06-25T01:17:46.391004Z","signature_status":"signed_v1","first_computed_at":"2026-06-25T01:17:46.391004Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Scale or Reason? A Compute-Equivalent Analysis of Reasoning Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"C\\'eline Hudelot, Hippolyte Gisserot-Boukhlef, Kevin El Haddad, Nicolas Boizard, Pierre Colombo","submitted_at":"2025-09-26T10:53:52Z","abstract_excerpt":"Distilling reasoning traces from strong teacher models has become the standard recipe for building capable small language models. Yet reasoning traces are 5-20$\\times$ longer than standard instruction fine-tuning (IFT) outputs, meaning every practitioner who chooses reasoning distillation implicitly forgoes training a larger IFT model on the same compute budget. Whether this trade-off is worthwhile remains unaddressed. We study it with a controlled experiment: a single teacher generates paired IFT and reasoning outputs for identical prompts by toggling only its reasoning mode, isolating superv"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.22193","kind":"arxiv","version":2},"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/2509.22193/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":"2509.22193","created_at":"2026-06-25T01:17:46.391062+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.22193v2","created_at":"2026-06-25T01:17:46.391062+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.22193","created_at":"2026-06-25T01:17:46.391062+00:00"},{"alias_kind":"pith_short_12","alias_value":"GSQZVRQPA6EV","created_at":"2026-06-25T01:17:46.391062+00:00"},{"alias_kind":"pith_short_16","alias_value":"GSQZVRQPA6EV3USS","created_at":"2026-06-25T01:17:46.391062+00:00"},{"alias_kind":"pith_short_8","alias_value":"GSQZVRQP","created_at":"2026-06-25T01:17:46.391062+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2606.29215","citing_title":"Multi-Block Diffusion Language Models","ref_index":29,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM","json":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM.json","graph_json":"https://pith.science/api/pith-number/GSQZVRQPA6EV3USSCVSHNUMRJM/graph.json","events_json":"https://pith.science/api/pith-number/GSQZVRQPA6EV3USSCVSHNUMRJM/events.json","paper":"https://pith.science/paper/GSQZVRQP"},"agent_actions":{"view_html":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM","download_json":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM.json","view_paper":"https://pith.science/paper/GSQZVRQP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.22193&json=true","fetch_graph":"https://pith.science/api/pith-number/GSQZVRQPA6EV3USSCVSHNUMRJM/graph.json","fetch_events":"https://pith.science/api/pith-number/GSQZVRQPA6EV3USSCVSHNUMRJM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM/action/storage_attestation","attest_author":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM/action/author_attestation","sign_citation":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM/action/citation_signature","submit_replication":"https://pith.science/pith/GSQZVRQPA6EV3USSCVSHNUMRJM/action/replication_record"}},"created_at":"2026-06-25T01:17:46.391062+00:00","updated_at":"2026-06-25T01:17:46.391062+00:00"}