让代码助手更聪明地提问，少问41%问题却能同样解决问题。
Asking What Matters: Reward-Driven Clarification for Software Engineering Tasks
- 用博弈论方法判断哪些信息关键，哪些问题用户能答。
- 训练出的模型比GPT-5少41%提问，但成功率相当。
- 适合开发工具、AI助手研发者参考实用优化策略。
人类在描述软件工程任务时常遗漏关键信息，因此助手需懂得何时以及如何提问以澄清需求。然而，有效澄清仍具挑战性，因为并非所有缺失信息都同等重要，且问题必须针对用户实际可回答的内容。我们通过量化分析真实软件任务中各类信息对任务成功的影响，并测试模拟用户对不同问题的响应能力，发现高效澄清依赖两个核心属性：任务相关性（何种信息能预测成功）与用户可回答性（用户能否提供）。我们将这两个属性转化为多阶段强化学习奖励，训练出一个80亿参数的澄清模块CLARITI。该模型在未充分定义的问题上达到与GPT-5相当的解决率，同时提问量减少41%。结果表明，基于信息影响与用户回答能力的实证分析，能显著提升澄清效率。
原文摘要 · Abstract (English)
Humans often specify tasks incompletely, so assistants must know when and how to ask clarifying questions. However, effective clarification remains challenging in software engineering tasks as not all missing information is equally valuable, and questions must target information users can realistically provide. We study clarification in real software engineering tasks by quantifying which types of information most affect task success and which questions elicit useful responses from simulated users. Using Shapley attribution and distributional comparisons, we identify two key properties of effective clarification: task relevance (which information predicts success) and user answerability (what users can realistically provide). We operationalize these properties as multi-stage reinforcement learning rewards to train CLARITI, an 8B-parameter clarification module, that matches GPT-5's resolution rate on underspecified issues while generating 41% fewer questions. Our results suggest that grounding reward design in empirical analysis of information impact and user answerability improves clarification efficiency.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。