用激活值估算置信度，让大模型更懂何时该沉默。
Confidence-Based Response Abstinence: Improving LLM Trustworthiness via Activation-Based Uncertainty Estimation
- 通过前馈层激活值建模置信度，避免概率信息损失。
- 在金融客服场景中，准确率高且响应延迟低。
- 适合对可靠性要求高的医疗、金融等关键领域。
我们提出一种检索增强生成（RAG）系统中的置信度估计方法，其与大语言模型输出的正确性高度一致。在金融、医疗等高风险领域，错误回答的成本远高于不回答。该方法利用前馈网络（FFN）原始激活值作为自回归信号，避免了经过投影和Softmax后导致的信息丢失，将置信度预测建模为序列分类任务，并采用Huber损失正则化训练以提升对噪声标注的鲁棒性。在真实金融行业客户支持场景中，面对复杂知识库，本方法优于强基线，在严格延迟约束下仍保持高准确率。在Llama 3.1 8B模型上实验表明，仅使用第16层激活值即可维持准确率并降低响应延迟。结果表明，基于激活值的置信度建模为可信RAG部署提供了可扩展、架构感知的路径。
原文摘要 · Abstract (English)
We propose a method for confidence estimation in retrieval-augmented generation (RAG) systems that aligns closely with the correctness of large language model (LLM) outputs. Confidence estimation is especially critical in high-stakes domains such as finance and healthcare, where the cost of an incorrect answer outweighs that of not answering the question. Our approach extends prior uncertainty quantification methods by leveraging raw feed-forward network (FFN) activations as auto-regressive signals, avoiding the information loss inherent in token logits and probabilities after projection and softmax normalization. We model confidence prediction as a sequence classification task, and regularize training with a Huber loss term to improve robustness against noisy supervision. Applied in a real-world financial industry customer-support setting with complex knowledge bases, our method outperforms strong baselines and maintains high accuracy under strict latency constraints. Experiments on Llama 3.1 8B model show that using activations from only the 16th layer preserves accuracy while reducing response latency. Our results demonstrate that activation-based confidence modeling offers a scalable, architecture-aware path toward trustworthy RAG deployment.
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。