{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ILAOHRUFCD7VLENSJ4AU6DK3OX","short_pith_number":"pith:ILAOHRUF","schema_version":"1.0","canonical_sha256":"42c0e3c68510ff5591b24f014f0d5b75c0117c6715c07ad3e28395739b0ca254","source":{"kind":"arxiv","id":"2505.11628","version":4},"attestation_state":"computed","paper":{"title":"Critique-Guided Distillation for Robust Reasoning via Refinement","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Berkcan Kapusuzoglu, Chia-Hsuan Lee, Sambit Sahu, Supriyo Chakraborty, Zain Sarwar","submitted_at":"2025-05-16T18:45:59Z","abstract_excerpt":"Supervised fine-tuning with expert demonstrations often produces models that imitate outputs without internalizing the reasoning processes needed for robust generalization. While critique-based approaches show promise, training models to generate critiques directly, such as Critique Fine-Tuning (CFT), can lead to output-format drift and degradation of general capabilities. We propose Critique-Guided Distillation (CGD), a training framework that decouples critique consumption from critique generation. During fine-tuning, the student is trained to refine flawed responses conditioned on teacher c"},"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":"2505.11628","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-16T18:45:59Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8cebaae1ea276260b3aebcb5a3b579af9da364db99afb09d50d0bc3ac9f626a9","abstract_canon_sha256":"af5d81425e307ab97e081c1b7c5ef1d2e424ab35994a0e160f9b73bac8672138"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T02:05:33.475720Z","signature_b64":"XtMmdqKzNtKndl0+7J9js9DlN55N44OqhxfbXGIJzHgwJRSwtfH4aB4S4vLfj1wlyr//zewZzhrOuwKuh3rYAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"42c0e3c68510ff5591b24f014f0d5b75c0117c6715c07ad3e28395739b0ca254","last_reissued_at":"2026-05-20T02:05:33.475042Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T02:05:33.475042Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Critique-Guided Distillation for Robust Reasoning via Refinement","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Berkcan Kapusuzoglu, Chia-Hsuan Lee, Sambit Sahu, Supriyo Chakraborty, Zain Sarwar","submitted_at":"2025-05-16T18:45:59Z","abstract_excerpt":"Supervised fine-tuning with expert demonstrations often produces models that imitate outputs without internalizing the reasoning processes needed for robust generalization. While critique-based approaches show promise, training models to generate critiques directly, such as Critique Fine-Tuning (CFT), can lead to output-format drift and degradation of general capabilities. We propose Critique-Guided Distillation (CGD), a training framework that decouples critique consumption from critique generation. During fine-tuning, the student is trained to refine flawed responses conditioned on teacher c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.11628","kind":"arxiv","version":4},"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/2505.11628/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":"2505.11628","created_at":"2026-05-20T02:05:33.475124+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.11628v4","created_at":"2026-05-20T02:05:33.475124+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.11628","created_at":"2026-05-20T02:05:33.475124+00:00"},{"alias_kind":"pith_short_12","alias_value":"ILAOHRUFCD7V","created_at":"2026-05-20T02:05:33.475124+00:00"},{"alias_kind":"pith_short_16","alias_value":"ILAOHRUFCD7VLENS","created_at":"2026-05-20T02:05:33.475124+00:00"},{"alias_kind":"pith_short_8","alias_value":"ILAOHRUF","created_at":"2026-05-20T02:05:33.475124+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX","json":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX.json","graph_json":"https://pith.science/api/pith-number/ILAOHRUFCD7VLENSJ4AU6DK3OX/graph.json","events_json":"https://pith.science/api/pith-number/ILAOHRUFCD7VLENSJ4AU6DK3OX/events.json","paper":"https://pith.science/paper/ILAOHRUF"},"agent_actions":{"view_html":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX","download_json":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX.json","view_paper":"https://pith.science/paper/ILAOHRUF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.11628&json=true","fetch_graph":"https://pith.science/api/pith-number/ILAOHRUFCD7VLENSJ4AU6DK3OX/graph.json","fetch_events":"https://pith.science/api/pith-number/ILAOHRUFCD7VLENSJ4AU6DK3OX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX/action/storage_attestation","attest_author":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX/action/author_attestation","sign_citation":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX/action/citation_signature","submit_replication":"https://pith.science/pith/ILAOHRUFCD7VLENSJ4AU6DK3OX/action/replication_record"}},"created_at":"2026-05-20T02:05:33.475124+00:00","updated_at":"2026-05-20T02:05:33.475124+00:00"}