How do LLMs Compute Verbal Confidence
Quick summary
arXiv:2603.17839v4 Announce Type: replace-cross Abstract: Verbal confidence -- prompting LLMs to state their confidence as a number or category -- is widely used to extract uncertainty estimates from black-box models. However, how LLMs internally generate such scores remains unknown. We address two questions: first, when confidence is computed -- just-in-time when requested, or automatically during answer generation and cached for later retrieval; and second, what verbal confidence represents -- token log-probabilities, or a richer evaluation of answer quality? Focusing on Gemma 3 27B (across
Key takeaways
- arXiv:2603.17839v4 Announce Type: replace-cross Abstract: Verbal confidence -- prompting LLMs to state their confidence as a number or category -- is widely used to extract uncertainty estimates from black-box models.
- However, how LLMs internally generate such scores remains unknown.
- We address two questions: first, when confidence is computed -- just-in-time when requested, or automatically during answer generation and cached for later retrieval; and second, what verbal confidence represents -- token log-probabilities, or a richer evaluation of answer quality?
Why it matters
“How do LLMs Compute Verbal Confidence” highlights the need for repeatable measurement rather than a single impressive demonstration. Independent validation across datasets and clearly stated limitations determine whether a result can guide product decisions.
Member comments