Reproduce “RebuttalAgent: Strategic Persuasion in Academic Rebuttal via Theory of Mind” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhitao He, Zongwei Lyu, Yi R Fung
RebuttalAgent, developed by researchers at the Hong Kong University of Science and Technology, integrates Theory of Mind (ToM) into Large Language Models (LLMs) to generate strategic and persuasive academic rebuttals. The framework achieved an average score of 9.42 on an automated evaluation and 9.57 in human evaluations, demonstrating a 7.36% improvement in persuasiveness over GPT-4.1.
Key insights
- Explicitly grounding academic rebuttal in Theory of Mind (ToM) principles transforms the task for LLMs from linguistic imitation into strategic reasoning and perspective-taking.
- A structured, multi-stage reasoning process (TSR framework) enables LLMs to systematically analyze reviewer profiles, plan specific strategies, and generate context-aware, evidence-based responses.
- Combining supervised fine-tuning with a novel self-reward mechanism during reinforcement learning allows for scalable self-improvement in strategic persuasion without relying on external, expert-annotated reward models.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.