RebuttalAgent-8B 📄🤝

An open 8-billion-parameter LLM that models a reviewer’s mindset (Theory of Mind) before writing, producing highly persuasive academic rebuttals.

---

1. Introduction

Although artificial intelligence (AI) has become deeply integrated into various stages of the research workflow and achieved remarkable advancements, academic rebuttal remains a significant and underexplored challenge. This is because rebuttal is a complex process of strategic communication under severe information asymmetry rather than a simple technical debate. Consequently, current approaches struggle as they largely imitate surface-level linguistics, missing the essential element of perspective-taking required for effective persuasion. In this paper, we introduce RebuttalAgent, the first framework to ground academic rebuttal in Theory of Mind (ToM), operationalized through a Theory-of-Mind-Strategy-Response (TSR) pipeline that models reviewer mental state, formulates persuasion strategy, and generates strategy-grounded response. To train our agent, we construct RebuttalBench, a large-scale dataset synthesized via a novel critique-and-refine approach. Our training process consists of two stages, beginning with a supervised fine-tuning phase to equip the agent with ToM-based analysis and strategic planning capabilities, followed by a reinforcement learning phase leveraging the self-reward mechanism for scalable self-improvement. For reliable and efficient automated evaluation, we further develop Rebuttal-RM, a specialized evaluator trained on over 100K samples of multi-source rebuttal data, which achieves scoring consistency with human preferences surpassing powerful judge GPT-4.1. Extensive experiments show RebuttalAgent significantly outperforms the base model by an average of 18.3% on automated metrics, while also proving with advanced proprietary models across both automated and human evaluations. The generated rebuttal content is for reference only, intended to inspire authors and assist in drafting. It cannot replace the author's own critical analysis and response.

---

2. Performance

2.1 Full results (0-10 scale, higher = better)

Model	Rigor			Soundness			Significance			Presentation			Avg
	C	P	Co	C	P	Co	C	P	Co	C	P	Co
o3	9.00	8.99	9.55	8.84	8.78	9.45	8.58	8.43	9.22	9.34	9.12	9.50	9.21
GPT-4.1	8.34	7.86	8.80	8.27	7.79	8.62	8.05	7.28	8.20	8.91	8.57	9.42	8.50
DeepSeek-R1	8.47	7.90	8.90	8.46	8.03	8.75	8.29	7.71	8.60	9.03	8.70	9.54	8.64
DeepSeek-V3	8.43	7.67	8.83	8.42	7.71	8.72	8.18	7.35	8.59	8.94	8.45	9.41	8.51
Gemini-2.5	7.89	6.91	6.63	8.06	7.41	7.26	7.87	7.09	6.89	8.56	8.11	8.83	7.75
GLM-4-9B	8.08	7.46	8.69	7.97	7.24	8.26	7.84	6.90	8.11	8.52	8.02	8.99	8.13
Llama-3.1-8B	7.77	6.69	7.32	7.71	6.76	7.02	7.54	6.30	6.49	8.12	7.42	8.25	7.44
Qwen-3-8B	7.96	7.33	8.18	7.84	7.11	7.76	7.68	6.73	7.39	8.51	8.08	8.87	7.96
Self-Refined	8.55	8.08	9.04	8.47	8.04	8.88	8.19	7.56	8.52	9.08	8.75	9.59	8.72
Strategy-Prompt	8.26	7.41	8.32	8.33	7.77	8.51	8.13	7.41	7.95	8.85	8.44	9.46	8.37
RebuttalFT	6.91	6.07	6.80	6.58	5.72	6.24	6.52	5.50	5.94	6.55	5.79	6.63	6.35
RebuttalAgent	9.23	8.91	9.59	9.18	8.95	9.37	9.09	8.54	9.65	9.43	9.20	9.50	9.42

Column abbreviations: C = Clarity, P = Persuasiveness, Co = Constructiveness.

Summary

As shown in Table 1, RebuttalAgent obtains the highest overall average score of 9.42, substantially outperforming all baselines, including GPT-4.1 and o3. It excels across the key rebuttal dimensions, achieving the best Clarity score (9.43) and a strong Persuasiveness score (9.20). Relative to the Qwen3-8B baseline, RebuttalAgent delivers an average improvement of 18.3 %, with the largest gains observed in Persuasiveness and Constructiveness (up to 34.6 %)

3 Deployment / Usage

3.1 Run with vLLM (OpenAI protocol)

pip install "vllm>=0.4.2"
python -m vllm.entrypoints.openai.api_server \
  --model Zhitao-He/RebuttalAgent \
  --dtype auto \
  --port 8000

3.2 Query via the official openai SDK

python

import openai
openai.api_key  = "EMPTY"
openai.base_url = "http://localhost:8000/v1"

prompt = f"""
The whole review content is{Full_Review_Content}
The target comment content is{Target_Comment}.
The best relevant paper fragment is{Relevant_Paper_Fragment}.
"""

resp = openai.chat.completions.create(
    model="Zhitao-He/RebuttalAgent",
    messages=[
        {"role": "system", "content": system_prompt}, 
        {"role": "user",   "content": user_prompt}
    ],
    temperature=0.7
)
print(resp.choices[0].message.content）

3.3 Default System Prompt (system_prompt)

 You are an expert academic assistant specializing in crafting persuasive and respectful rebuttals for peer reviews. Your goal is to formulate a response that addresses the reviewer's concerns directly and constructively, ultimately strengthening the paper's position for acceptance. You receive the following inputs:

Full_Review_Content: The entire review text for the target paper.
Target_Comment: A specific excerpt from the review that requires a response.
Relevant_Paper_Fragment: A key excerpt from the author's own manuscript. This fragment provides the essential context and technical details that relevant to the Target_Comment.

Your task is to generate a structured rebuttal plan and response by following these steps precisely:

Step 1: Analysis 
First, conduct your analysis of the overall review and target comment. Present this analysis inside <analysis> and </analysis> tags using the strict JSON format specified below. 

Step 2: Rebuttal Strategy 
Based on your analysis and the information within the Relevant_Paper_Fragment, devise an optimal, step-by-step strategy for the response. Present this strategy as a numbered list inside <strategy> and </strategy> tags. Each step should be a clear action.
Step 3: Rebuttal Response 
Finally, craft the rebuttal response for the Target_Comment. Write the response inside <response> and </response>, based on your above analysis and strategy.

Here is an example of output format:

I need to analysis the review’s overall instance and the target comment: 
<analysis>
{
  "global_profile": {
    "overall_stance": "...",
    "overall_attitude": "...",
    "dominant_concern": "...",
    "reviewer_expertise": "..."
  },
  "comment_analysis": 
    {
      "comment_text": "...",
      "category": "...",
      "sub_category": "...",
      "severity": "..."
}
}
</analysis>.
Based on current overall analysis, to address the target comment, I need to adopt the following strategies:
<strategy> 1. ; 2. ; 3. ; XXX</strategy>.
Based on the above analysis and strategies, for the target comment: <response>XXX</response>.

4 Citation

@misc{he2026dancingchainsstrategicpersuasion,
      title={Dancing in Chains: Strategic Persuasion in Academic Rebuttal via Theory of Mind}, 
      author={Zhitao He and Zongwei Lyu and Yi R Fung},
      year={2026},
      eprint={2601.15715},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2601.15715}, 
}