{"ID":5675215,"CreatedAt":"2026-07-03T01:40:09.565152011Z","UpdatedAt":"2026-07-05T07:54:18.289289986Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.01846","arxiv_id":"2607.01846","title":"CLAP: Closed-Loop Training, Evaluation, and Release Control for Domain Agent Post-training","abstract":"Domain agents often face noisy business data, uncertain post-training gains, offline/application mismatch, and adapter-release risk. This paper presents CLAP (Closed-Loop Agent Post-training), a closed-loop method that converts business data into structured SFT samples, decision-preference samples, holdout sets, risk diagnostics, and release-gate records. CLAP combines data validation, target/evidence normalization, reward/KL diagnosis, offline gates, and application-chain replay to decide whether an adapter is suitable for the target application chain. On five anonymized manufacturing-scenario batches, QLoRA-style LoRA-SFT yields modest average gains: overall score increases by 0.0098, pass rate by 0.0240, and evidence accuracy by 0.0280, while hallucination and wrong facts decrease. Yet only 3 of 5 batches improve, some batches regress, and GRPO exposes high KL risks. Application-chain replay further shows that RAG is necessary for factual extraction; under the same 3B backbone and 100 replay cases, an application-RAG-oriented LoRA-SFT adapter improves value, core fields, and answer-evidence doc/page matching over base+RAG, but increases latency. These results support managing domain-agent post-training through an integrated data-training-evaluation-release loop rather than relying on training completion or a single offline score.","short_abstract":"Domain agents often face noisy business data, uncertain post-training gains, offline/application mismatch, and adapter-release risk. This paper presents CLAP (Closed-Loop Agent Post-training), a closed-loop method that converts business data into structured SFT samples, decision-preference samples, holdout sets, risk d...","url_abs":"https://arxiv.org/abs/2607.01846","url_pdf":"https://arxiv.org/pdf/2607.01846v1","authors":"[\"Fangfei Li\",\"Chenyang Zhao\",\"Long Wang\",\"Feng Tian\",\"Zhiyue Zheng\",\"Lv Guo\"]","published":"2026-07-02T08:07:56Z","proceeding":"cs.AI","tasks":"[\"cs.AI\"]","methods":"[\"LoRA\"]","has_code":false}