Uncertainty Distillation: Teaching Language Models to Express Semantic Confidence
원문 초록
As large language models (LLMs) are increasingly used for factual question-answering, it becomes more important for LLMs to have the capability to communicate the likelihood that their answer is correct. For these verbalized expressions of uncertainty to be meaningful, they should reflect the error rates at the expressed level of confidence. However, when prompted to express confidence, the error rates of current LLMs are inconsistent with their communicated confidences, highlighting the need for uncertainty quantification methods. Many prior methods calculate lexical uncertainty, estimating a model's confidence in the specific string it generated. In some cases, however, it may be more useful to estimate semantic uncertainty, or the model's confidence in the answer regardless of how it is verbalized. We propose a simple procedure, uncertainty distillation, to teach an LLM to verbalize calibrated semantic confidences. Using held-out data to map initial uncertainty estimates to meaningful probabilities, we create examples annotated with verbalized probabilities for supervised fine-tuning. We compare uncertainty distillation to several strong baselines, and find that our method yields verbalized confidences that correlate well with observed error rates.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
UProp: Investigating the Uncertainty Propagation of LLMs in Multi-Step Agentic Decision-Making
COIN: Uncertainty-Guarding Selective Question Answering for Foundation Models with Provable Risk Guarantees
A Fast, Reliable, and Secure Programming Language for LLM Agents with Code Actions
Statistical Machine Learning for Astronomy -- A Textbook
From Roots to Rewards: Dynamic Tree Reasoning with RL
Describe Anything Model for Visual Question Answering on Text-rich Images
토론
이 논문과 연결된 포럼 글이 아직 없습니다.