发表机构
Yale University; University of Chicago; TCS Research（耶鲁大学; 芝加哥大学; 塔塔咨询服务研究中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对大模型预投稿自我审查中反馈缺乏可操作性的问题，提出ActReview框架，利用反驳引导构建训练数据并设计评分奖励，生成诊断与修改建议，优于现有模型。
AI 中文摘要
随着大型语言模型越来越多地用于投稿前的自我审查，人们对反馈的需求日益增长，这种反馈不仅要指出弱点，还要引导作者进行具体的修改。我们将此研究为可操作的同行评审生成，并将其分解为两个子任务：诊断性声明生成和修改建议生成。我们引入了ActReview，一个基于反驳引导的后训练框架，将针对论文的具体诊断与具体且有理有据的修改计划联系起来。我们的核心见解是，作者的反驳揭示了解决评审人关切的可操作行动，因此可以为面向修改的反馈提供潜在监督。从OpenReview上的真实评审-反驳线程中，我们通过将评审人的弱点与作者的回应对齐，并将由此产生的反馈基于局部论文证据，构建了ActReview-40K。我们使用多任务监督微调对Qwen3-8B-Base进行后训练，随后使用基于候选感知、弱点特定评分标准的GRPO。我们还引入了ActReview-Bench，一个包含1,000个实例的人工策划基准，用于评估诊断质量和修改有用性。实验表明，ActReview在可操作性和基于证据方面优于先前专门的评审生成模型，同时与基于强提示的LLM保持竞争力。人工评估证实了修改有用性的提升，同时揭示了在技术准确性方面仍存在的差距，进一步的分析支持对未见论文的泛化性和跨独立评审者的稳健性。
英文摘要
As LLMs are increasingly used for pre-submission self-review, there is growing demand for feedback that not only identifies weaknesses but also guides authors toward concrete revisions. We study this as Actionable Peer-review Generation and decompose it into two subtasks: diagnostic claim generation and revision suggestion generation. We introduce ActReview, a rebuttal-guided post-training framework that connects paper-specific diagnoses to concrete, grounded revision plans. Our central insight is that author rebuttals reveal plausible actions for addressing reviewer concerns and can therefore provide latent supervision for revision-oriented feedback. From real review-rebuttal threads on OpenReview, we construct ActReview-40K by aligning reviewer weaknesses with author responses and grounding the resulting feedback in localized paper evidence. We post-train Qwen3-8B-Base with multi-task supervised fine-tuning followed by GRPO using candidate-aware, weakness-specific rubric rewards. We also introduce ActReview-Bench, a human-curated benchmark of 1,000 instances for evaluating diagnostic quality and revision usefulness. Experiments show that ActReview outperforms prior specialized review-generation models on actionability and grounding while remaining competitive with strong prompt-based LLMs. Human evaluation confirms improved revision usefulness while revealing a remaining gap in technical accuracy, and additional analyses support generalization to held-out papers and robustness across independent judges.
Comments50 pages, 20 figures
