用无答案背景信息重写问题，让模型答题准确率翻倍
Query Disambiguation via Answer-Free Context: Doubling Performance on Humanity's Last Exam
- 用无答案上下文动态构建背景并重写问题，降低歧义
- 在人类最后测验上，GPT-5-mini准确率从14%提升至37%
- 需先重写再回答，仅靠推理提示无法达到同样效果
问题表述的清晰程度对语言模型和人类的回答质量均有深远影响。尽管模型能力持续提升，但上下文与问题表述之间的相互作用仍研究不足。本文研究了模型上下文窗口中背景信息的质量对准确性的影响。发现将良好接地的动态上下文构建（即RAG）与问题重写相结合，可显著降低问题歧义，带来准确率提升。给定一个用户问题及其无答案的背景上下文，通过重写问题以减少歧义，可在不改变答案的前提下实现基准性能提升，甚至优于直接将上下文前置到问题前。使用gpt-oss-20b重写Humanity's Last Exam的部分问题，使gpt-5-mini的准确率从0.14提升至0.37。我们证明，这种提升无法仅通过推理时的提示恢复；必须分重写与回答两个阶段。代码与数据已开源。
原文摘要 · Abstract (English)
How carefully and unambiguously a question is phrased has a profound impact on the quality of the response, for Language Models (LMs) as well as people. While model capabilities continue to advance, the interplay between grounding context and query formulation remains under-explored. This work investigates how the quality of background grounding information in a model's context window affects accuracy. We find that combining well-grounded dynamic context construction (i.e, RAG) with query rewriting reduces question ambiguity, resulting in significant accuracy gains. Given a user question with associated answer-free grounding context, rewriting the question to reduce ambiguity produces benchmark improvements without changing the answer itself, even compared to prepending that context before the question. Using \texttt{gpt-oss-20b} to rewrite a subset of Humanity's Last Exam using answer-free grounding context improves \texttt{gpt-5-mini} accuracy from 0.14 to 0.37. We demonstrate that this accuracy improvement cannot be fully recovered just through prompting at inference time; rather, distinct rewriting and answering phases are required. Code and data are available at https://github.com/mmajurski/lm-rewrite-uplift
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。