{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:W4UO4CJZI6HA3YJ7B3M4PNHM5V","short_pith_number":"pith:W4UO4CJZ","schema_version":"1.0","canonical_sha256":"b728ee0939478e0de13f0ed9c7b4eced6117a166a3bf07cbcb1c362780079d9b","source":{"kind":"arxiv","id":"2409.19715","version":2},"attestation_state":"computed","paper":{"title":"Coffee-Gym: An Environment for Evaluating and Improving Natural Language Feedback on Erroneous Code","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Beong-woo Kwak, Dongjin Kang, Hyungjoo Chae, Jinyoung Yeo, Kai Tzu-iunn Ong, Seonghyeon Bae, Seungjun Moon, Seung-won Hwang, Taeyoon Kwon, Yongho Song","submitted_at":"2024-09-29T14:14:25Z","abstract_excerpt":"This paper presents Coffee-Gym, a comprehensive RL environment for training models that provide feedback on code editing. Coffee-Gym includes two major components: (1) Coffee, a dataset containing humans' code edit traces for coding questions and machine-written feedback for editing erroneous code; (2) CoffeeEval, a reward function that faithfully reflects the helpfulness of feedback by assessing the performance of the revised code in unit tests. With them, Coffee-Gym addresses the unavailability of high-quality datasets for training feedback models with RL, and provides more accurate rewards "},"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":"2409.19715","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-29T14:14:25Z","cross_cats_sorted":[],"title_canon_sha256":"fbf30f9c60a7a97a5c1dd72835fbbdd172388fc59e45ad2a134db6fa0df9c86c","abstract_canon_sha256":"809787940014f4c32798a45e0ef8e95b75e9c99f7f53876a4d44f0b28cc06ae6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:15:38.279243Z","signature_b64":"+uQVslCFRJ9fPaJDtPkRQdt2D5mohLO94T4u53a872yteIcomGWtBqIuxfuuHK+zRlEMS0+kgf5w8ffpiMgxDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b728ee0939478e0de13f0ed9c7b4eced6117a166a3bf07cbcb1c362780079d9b","last_reissued_at":"2026-07-05T09:15:38.278767Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:15:38.278767Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Coffee-Gym: An Environment for Evaluating and Improving Natural Language Feedback on Erroneous Code","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Beong-woo Kwak, Dongjin Kang, Hyungjoo Chae, Jinyoung Yeo, Kai Tzu-iunn Ong, Seonghyeon Bae, Seungjun Moon, Seung-won Hwang, Taeyoon Kwon, Yongho Song","submitted_at":"2024-09-29T14:14:25Z","abstract_excerpt":"This paper presents Coffee-Gym, a comprehensive RL environment for training models that provide feedback on code editing. Coffee-Gym includes two major components: (1) Coffee, a dataset containing humans' code edit traces for coding questions and machine-written feedback for editing erroneous code; (2) CoffeeEval, a reward function that faithfully reflects the helpfulness of feedback by assessing the performance of the revised code in unit tests. With them, Coffee-Gym addresses the unavailability of high-quality datasets for training feedback models with RL, and provides more accurate rewards "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.19715","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/2409.19715/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":"2409.19715","created_at":"2026-07-05T09:15:38.278824+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.19715v2","created_at":"2026-07-05T09:15:38.278824+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.19715","created_at":"2026-07-05T09:15:38.278824+00:00"},{"alias_kind":"pith_short_12","alias_value":"W4UO4CJZI6HA","created_at":"2026-07-05T09:15:38.278824+00:00"},{"alias_kind":"pith_short_16","alias_value":"W4UO4CJZI6HA3YJ7","created_at":"2026-07-05T09:15:38.278824+00:00"},{"alias_kind":"pith_short_8","alias_value":"W4UO4CJZ","created_at":"2026-07-05T09:15:38.278824+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.16667","citing_title":"ELABORATION: A Comprehensive Benchmark on Human-LLM Competitive Programming","ref_index":2024,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V","json":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V.json","graph_json":"https://pith.science/api/pith-number/W4UO4CJZI6HA3YJ7B3M4PNHM5V/graph.json","events_json":"https://pith.science/api/pith-number/W4UO4CJZI6HA3YJ7B3M4PNHM5V/events.json","paper":"https://pith.science/paper/W4UO4CJZ"},"agent_actions":{"view_html":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V","download_json":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V.json","view_paper":"https://pith.science/paper/W4UO4CJZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.19715&json=true","fetch_graph":"https://pith.science/api/pith-number/W4UO4CJZI6HA3YJ7B3M4PNHM5V/graph.json","fetch_events":"https://pith.science/api/pith-number/W4UO4CJZI6HA3YJ7B3M4PNHM5V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V/action/storage_attestation","attest_author":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V/action/author_attestation","sign_citation":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V/action/citation_signature","submit_replication":"https://pith.science/pith/W4UO4CJZI6HA3YJ7B3M4PNHM5V/action/replication_record"}},"created_at":"2026-07-05T09:15:38.278824+00:00","updated_at":"2026-07-05T09:15:38.278824+00:00"}