This research proposes a novel approach to uncertainty quantification (UQ) in large language models by examining multiple knowledge dimensions within responses.

• Introduces a multi-dimensional framework for assessing LLM response reliability
• Evaluates uncertainty across factual accuracy, logical coherence, and content relevance dimensions
• Provides more comprehensive reliability metrics than traditional semantic similarity methods
• Demonstrates improved uncertainty detection in high-stakes domains like healthcare

For medical applications, this approach significantly enhances safety by identifying when LLMs may provide unreliable information for clinical decision-making, reducing potential harm from AI-assisted healthcare systems.

Uncertainty Quantification of Large Language Models through Multi-Dimensional Responses