ReFeed rethinks summary refinement by showing that reflective reasoning over multi-dimensional feedback is key to achieving balanced, robust improvements.
We recommend using the following prompt. For details on the feedback format and the code, please refer to
our GitHub repository.
1
2system_prompt = "Your role as an assistant involves thoroughly exploring questions through a systematic long thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracing, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Think and Answer. In the Think section, detail your reasoning process using the specified format: <think> {thought with steps separated with '\\n\\n'} </think> Each step should include detailed considerations such as analisying questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Answer section, based on various attempts, explorations, and reflections from the Think section, systematically present the final solution that you deem correct. The solution should remain a logical, accurate, concise expression style and detail necessary step needed to reach the conclusion, formatted as follows: <answer> {final formatted, precise, and clear solution} </answer> Now, try to solve the following question through the above guidelines:"
3
4user_prompt = f"""
5Your goal is to deliberate on the provided feedback and propose actionable and specific aggregated feedback based on it.
6
7Instructions:
81. Deliberate on the characteristics an ideal summary should achieve.
92. Assess and choose the validity of the given feedback in improving the summary considering feedback quality criteria:
10- Faithfulness: Out-of-article, Entity, Relation, Sentence errors
11- Completeness: missing key content
12- Conciseness: unnecessary or non-key content
133. Aggregate the valid feedback and revise the summary by incorporating it.
144. Check whether revisions harm other quality dimensions.
15
16Document:
17{document}
18
19Summary:
20{summary}
21
22Feedback:
23{feedback}
24"""
25
Please consider citation if our paper is useful in your research.
1@inproceedings{yun2025refeed,
2 title={ReFeed: Multi-dimensional Summarization Refinement with Reflective Reasoning on Feedback},
3 author={Taewon Yun and Jihwan Oh and Hyangsuk Min and Yuho Lee and Jihwan Bang and Jason Cai and Hwanjun Song},
4 booktitle={Second Conference on Language Modeling},
5 year={2025},
6 url={https://openreview.net/forum?id=6BGDGKZN7q}
7}
This research was supported by KISTI, and by the NRF. For GPU infrastructure, our work was supported by the IITP grant funded by MSIT.
This work was supported by Institute of Information & Communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (RS-2025-02653113, High-Performance Research AI Computing Infrastructure Support at the 2 PFLOPS Scale)
