通过双重校准提升大模型推理可靠性，降低幻觉风险。
Double-Calibration: Towards Reliable LLMs via Calibrating Knowledge and Reasoning Confidence
- 用轻量代理模型生成带置信度的知识图谱证据
- 使大模型输出更准确且置信度与真实可信度匹配
- 适合需要高可靠性的问答与决策系统
大语言模型在推理中易产生幻觉，影响可靠性。尽管引入知识图谱可提升事实准确性，但现有方法无法量化检索证据和模型推理中的认知不确定性。为此，我们提出双校准框架DoublyCal，基于新颖的双重校准原则：首先使用轻量代理模型生成带校准置信度的知识图谱证据；再以此证据指导黑箱大模型，生成最终预测。结果不仅更准确，且置信度可追溯至支持证据的不确定性。在知识密集型基准测试中，DoublyCal显著提升黑箱大模型的准确率与置信度校准效果，同时保持低令牌开销。
原文摘要 · Abstract (English)
Reliable reasoning in Large Language Models (LLMs) is challenged by their propensity for hallucination. While augmenting LLMs with Knowledge Graphs (KGs) improves factual accuracy, existing KG-augmented methods fail to quantify epistemic uncertainty in both the retrieved evidence and LLMs' reasoning. To bridge this gap, we introduce DoublyCal, a framework built on a novel double-calibration principle. DoublyCal employs a lightweight proxy model to first generate KG evidence alongside a calibrated evidence confidence. This calibrated supporting evidence then guides a black-box LLM, yielding final predictions that are not only more accurate but also well-calibrated, with confidence scores traceable to the uncertainty of the supporting evidence. Experiments on knowledge-intensive benchmarks show that DoublyCal significantly improves both the accuracy and confidence calibration of black-box LLMs while maintaining low token cost.
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。