提出轻量框架HalluClean，自动识别并修正大模型幻觉内容。
HalluClean: A Unified Framework to Combat Hallucinations in LLMs
- 分规划、执行、修订三阶段推理，精准定位幻觉
- 零样本跨任务通用，无需外部知识或标注数据
- 在5类任务中显著提升事实一致性，适合实际应用
大语言模型在多项自然语言处理任务中表现优异，但常生成缺乏事实依据的幻觉内容。为此，我们提出HalluClean，一种轻量级、任务无关的幻觉检测与修正框架。该框架采用增强推理范式，将过程显式分解为规划、执行和修订三个阶段，以识别并修正无支持的陈述。通过极简的任务路由提示，实现零样本跨域泛化，不依赖外部知识源或监督检测器。我们在五个代表性任务——问答、对话、摘要、数学应用题和矛盾检测上进行了全面评估。实验结果表明，HalluClean显著提升了事实一致性，优于多个竞争性基线，展现了其在真实场景中增强大模型输出可信度的潜力。
原文摘要 · Abstract (English)
Large language models (LLMs) have achieved impressive performance across a wide range of natural language processing tasks, yet they often produce hallucinated content that undermines factual reliability. To address this challenge, we introduce HalluClean, a lightweight and task-agnostic framework for detecting and correcting hallucinations in LLM-generated text. HalluClean adopts a reasoning-enhanced paradigm, explicitly decomposing the process into planning, execution, and revision stages to identify and refine unsupported claims. It employs minimal task-routing prompts to enable zero-shot generalization across diverse domains, without relying on external knowledge sources or supervised detectors. We conduct extensive evaluations on five representative tasks-question answering, dialogue, summarization, math word problems, and contradiction detection. Experimental results show that HalluClean significantly improves factual consistency and outperforms competitive baselines, demonstrating its potential to enhance the trustworthiness of LLM outputs in real-world applications.
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。