Language Models (Mostly) Know What They Know

Abstract

We study whether language models can evaluate the validity of their own claims and predict which questions they will be able to answer correctly. We first show that larger models are well-calibrated on diverse multiple choice and true/false questions when they are provided in the right format. Thus we can approach self-evaluation on open-ended sampling tasks by asking models to first propose answers, and then to evaluate the probability "P(True)" that their answers are correct. We find encouraging performance, calibration, and scaling for P(True) on a diverse array of tasks. Performance at self-evaluation further improves when we allow models to consider many of their own samples before predicting the validity of one specific possibility. Next, we investigate whether models can be trained to predict "P(IK)", the probability that "I know" the answer to a question, without reference to any particular proposed answer. Models perform well at predicting P(IK) and partially generalize across tasks, though they struggle with calibration of P(IK) on new tasks. The predicted P(IK) probabilities also increase appropriately in the presence of relevant source materials in the context, and in the presence of hints towards the solution of mathematical word problems. We hope these observations lay the groundwork for training more honest models, and for investigating how honesty generalizes to cases where models are trained on objectives other than the imitation of human writing.

arXiv comment23+17 pages; refs added, typos fixed

Cited by in corpus (11)

• A Survey on Large Language Model (LLM) Security and Privacy: The Good, the Bad, and the Ugly
• Large Legal Fictions: Profiling Legal Hallucinations in Large Language Models
• CoAnnotating: Uncertainty-Guided Work Allocation between Human and Large Language Models for Data Annotation
• Automated Review Generation Method Based on Large Language Models
• AI-Driven Day-to-Day Route Choice
• Benchmarking Uncertainty Quantification Methods for Large Language Models with LM-Polygraph
• Transforming Agency. On the mode of existence of Large Language Models
• GistVis: Automatic Generation of Word-scale Visualizations from Data-rich Documents
• Surfacing Variations to Calibrate Perceived Reliability of MLLM-generated Image Descriptions
• On the attribution of confidence to large language models
• Visual hallucination detection in large vision-language models via evidential conflict