GrACE让大模型实时生成可靠置信度，提升安全性和推理效率。
GrACE: A Generative Approach to Better Confidence Elicitation and Efficient Test-Time Scaling in Large Language Models
- 通过特殊标记嵌入与隐状态相似度计算置信度，实时生成。
- 置信度校准效果优于现有方法，无需额外采样或辅助模型。
- 支持测试时缩放策略，显著减少采样次数，适合高风险场景。
在医疗、金融等高风险应用中，通过置信度评估来衡量大语言模型（LLM）的可靠性是AI安全的重要方向。现有方法或计算开销大，或校准效果差，难以实际部署。本文提出GrACE——一种生成式置信度提取方法，使LLM能实现可扩展且可靠的置信度估计。GrACE采用新机制：模型通过最后隐藏状态与词汇表中特殊标记嵌入的相似度，在实时生成中表达置信度。我们对模型进行微调，以准确率相关目标校准置信度。大量实验表明，GrACE在开放式生成任务中实现了最佳判别能力与校准性能，无需额外采样或辅助模型。此外，我们提出了两种基于置信度的测试时缩放策略，不仅提升了最终决策准确性，还显著减少了所需样本数，凸显其在部署中实现可靠、即时置信度估计的潜力。
原文摘要 · Abstract (English)
Assessing the reliability of Large Language Models (LLMs) by confidence elicitation is a prominent approach to AI safety in high-stakes applications, such as healthcare and finance. Existing methods either require expensive computational overhead or suffer from poor calibration, making them impractical and unreliable for real-world deployment. In this work, we propose GrACE, a Generative Approach to Confidence Elicitation that enables scalable and reliable confidence elicitation for LLMs. GrACE adopts a novel mechanism in which the model expresses confidence by the similarity between the last hidden state and the embedding of a special token appended to the vocabulary, in real-time. We fine-tune the model for calibrating the confidence with targets associated with accuracy. Extensive experiments show that the confidence produced by GrACE achieves the best discriminative capacity and calibration on open-ended generation tasks without resorting to additional sampling or an auxiliary model. Moreover, we propose two confidence-based strategies for test-time scaling with GrACE, which not only improve the accuracy of the final decision but also significantly reduce the number of required samples, highlighting its potential as a practical solution for deploying LLMs with reliable, on-the-fly confidence estimation.
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。