How do LLMs Compute Verbal Confidence

1. 1Investigate the internal mechanisms of LLMs to determine when verbal confidence is computed
2. 2Analyze the relationship between answer generation and confidence calculation to understand if confidence is calculated just-in-time or cached
3. 3Examine the representation of verbal confidence to determine what it signifies in the context of LLMs
4. 4Evaluate the implications of verbal confidence computation on model performance and uncertainty estimation

AI engineers and researchers benefit from understanding how LLMs generate uncertainty estimates, which can inform model development and improvement

💡 Verbal confidence computation in LLMs is not well understood, but research can uncover its internal mechanisms and representation

🤖 How do LLMs compute verbal confidence? New research sheds light on internal mechanisms 📊