{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TXN65WMTBKD7CBIEVL6GHOXCU7","short_pith_number":"pith:TXN65WMT","schema_version":"1.0","canonical_sha256":"9ddbeed9930a87f10504aafc63bae2a7ed66f0cce57658020def5fa528dd25f6","source":{"kind":"arxiv","id":"2606.00020","version":1},"attestation_state":"computed","paper":{"title":"CSRP: Chain-of-Thought Reasoning for Chinese Text Correction via Reinforcement Learning with Efficiency-Aware Rewards","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Man Lan, Wei Tian, Yuhao Zhou","submitted_at":"2026-04-14T02:06:38Z","abstract_excerpt":"Large Language Model (LLM) based Chinese Grammatical Error Correction (CGEC) systems face two critical challenges: general-purpose models lack specialized linguistic priors for subtle grammatical distinctions, and Supervised Fine-Tuning (SFT) with Maximum Likelihood Estimation fails to optimize for precision-focused metrics, leading to systematic over-correction. We propose CSRP, a three-stage framework that progressively builds correction capability through Continual Pre-training (CPT) on 5.9M balanced samples to internalize domain knowledge, Chain-of-Thought SFT with explicit error reasoning"},"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":"2606.00020","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-04-14T02:06:38Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"2955f5668cf29380a9ce3555aee0ae3d730f5ae80f76ad190343a09305e0f293","abstract_canon_sha256":"ff3101135cd528e946037c440da3f339cb51f9020f79f2e4aaf4a71a11c20285"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T00:03:12.546498Z","signature_b64":"6PHXwcOKTVWw+wSwuwKe0Q/HmcbIbe/l3+GpsTcFXwJtPrjhUMr/w4/0iUreSfzGMbD+7UIEqdljzXYtvrsTCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9ddbeed9930a87f10504aafc63bae2a7ed66f0cce57658020def5fa528dd25f6","last_reissued_at":"2026-06-02T00:03:12.546026Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T00:03:12.546026Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CSRP: Chain-of-Thought Reasoning for Chinese Text Correction via Reinforcement Learning with Efficiency-Aware Rewards","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Man Lan, Wei Tian, Yuhao Zhou","submitted_at":"2026-04-14T02:06:38Z","abstract_excerpt":"Large Language Model (LLM) based Chinese Grammatical Error Correction (CGEC) systems face two critical challenges: general-purpose models lack specialized linguistic priors for subtle grammatical distinctions, and Supervised Fine-Tuning (SFT) with Maximum Likelihood Estimation fails to optimize for precision-focused metrics, leading to systematic over-correction. We propose CSRP, a three-stage framework that progressively builds correction capability through Continual Pre-training (CPT) on 5.9M balanced samples to internalize domain knowledge, Chain-of-Thought SFT with explicit error reasoning"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00020","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/2606.00020/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":"2606.00020","created_at":"2026-06-02T00:03:12.546089+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00020v1","created_at":"2026-06-02T00:03:12.546089+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00020","created_at":"2026-06-02T00:03:12.546089+00:00"},{"alias_kind":"pith_short_12","alias_value":"TXN65WMTBKD7","created_at":"2026-06-02T00:03:12.546089+00:00"},{"alias_kind":"pith_short_16","alias_value":"TXN65WMTBKD7CBIE","created_at":"2026-06-02T00:03:12.546089+00:00"},{"alias_kind":"pith_short_8","alias_value":"TXN65WMT","created_at":"2026-06-02T00:03:12.546089+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/TXN65WMTBKD7CBIEVL6GHOXCU7","json":"https://pith.science/pith/TXN65WMTBKD7CBIEVL6GHOXCU7.json","graph_json":"https://pith.science/api/pith-number/TXN65WMTBKD7CBIEVL6GHOXCU7/graph.json","events_json":"https://pith.science/api/pith-number/TXN65WMTBKD7CBIEVL6GHOXCU7/events.json","paper":"https://pith.science/paper/TXN65WMT"},"agent_actions":{"view_html":"https://pith.science/pith/TXN65WMTBKD7CBIEVL6GHOXCU7","download_json":"https://pith.science/pith/TXN65WMTBKD7CBIEVL6GHOXCU7.json","view_paper":"https://pith.science/paper/TXN65WMT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00020&json=true","fetch_graph":"https://pith.science/api/pith-number/TXN65WMTBKD7CBIEVL6GHOXCU7/graph.json","fetch_events":"https://pith.science/api/pith-number/TXN65WMTBKD7CBIEVL6GHOXCU7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TXN65WMTBKD7CBIEVL6GHOXCU7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TXN65WMTBKD7CBIEVL6GHOXCU7/action/storage_attestation","attest_author":"https://pith.science/pith/TXN65WMTBKD7CBIEVL6GHOXCU7/action/author_attestation","sign_citation":"https://pith.science/pith/TXN65WMTBKD7CBIEVL6GHOXCU7/action/citation_signature","submit_replication":"https://pith.science/pith/TXN65WMTBKD7CBIEVL6GHOXCU7/action/replication_record"}},"created_at":"2026-06-02T00:03:12.546089+00:00","updated_at":"2026-06-02T00:03:12.546089+00:00"}