# Confidence in the Reasoning of Large Language Models

arXiv: https://arxiv.org/abs/2412.15296 · HTML: https://mlsanity.com/paper/2412.15296/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- First posted on arXiv 19 Dec 2024.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Built on by (most cited first, 18 of 18)

- [No Answer Needed: Predicting LLM Answer Accuracy from Question-Only Linear Probes](https://mlsanity.com/paper/2509.10625/) (2025)
- [BrokenMath: A Benchmark for Sycophancy in Theorem Proving with LLMs](https://mlsanity.com/paper/2510.04721/) (2025)
- [Towards Automating Data Access Permissions in AI Agents](https://mlsanity.com/paper/2511.17959/) (2025)
- [Mitigating LLM Hallucination via Behaviorally Calibrated Reinforcement Learning](https://mlsanity.com/paper/2512.19920/) (2025)
- [Read Your Own Mind: Reasoning Helps Surface Self-Confidence Signals in LLMs](https://mlsanity.com/paper/2505.23845/) (2025)
- [Agentic LLM Framework for Adaptive Decision Discourse](https://mlsanity.com/paper/2502.10978/) (2025)
- [Martingale Score: An Unsupervised Metric for Bayesian Rationality in LLM Reasoning](https://mlsanity.com/paper/2512.02914/) (2025)
- [Lightweight Latent Verifiers for Efficient Meta-Generation Strategies](https://mlsanity.com/paper/2504.16760/) (2025)
- [Bi-directional Model Cascading with Proxy Confidence](https://mlsanity.com/paper/2504.19391/) (2025)
- [A Real-World Evaluation of LLM Medication Safety Reviews in NHS Primary Care](https://mlsanity.com/paper/2512.21127/) (2025)
- [TruthTensor: Evaluating LLMs through Human Imitation on Prediction Market under Drift and Holistic Reasoning](https://mlsanity.com/paper/2601.13545/) (2026)
- [PrivScope: Task-scoped Disclosure Control for Hybrid Agentic Systems](https://mlsanity.com/paper/2605.16630/) (2026)
- [Beyond Misinformation: A Conceptual Framework for Studying AI Hallucinations in (Science) Communication](https://mlsanity.com/paper/2504.13777/) (2025)
- [Improving MLLM Historical Record Extraction with Test-Time Image](https://mlsanity.com/paper/2509.09722/) (2025)
- [The Dunning-Kruger Effect in Large Language Models: An Empirical Study of Confidence Calibration](https://mlsanity.com/paper/2603.09985/) (2026)
- [On the Limits of LLM Adaptability: Impact of Model-Internalized Priors on Annotation Task Performance](https://mlsanity.com/paper/2606.00467/) (2026)
- [Identifying High-Confidence Social Biases in LLMs for Trustworthy Conversational Tutoring Agents](https://mlsanity.com/paper/2606.01584/) (2026)
- [Distilling What Matters: Confidence-Aware Selective Distillation for Large Language Models](https://mlsanity.com/paper/2609.36734/) (2026)

## Abstract (arXiv)

There is a growing literature on reasoning by large language models (LLMs), but the discussion on the uncertainty in their responses is still lacking. Our aim is to assess the extent of confidence that LLMs have in their answers and how it correlates with accuracy. Confidence is measured (i) qualitatively in terms of persistence in keeping their answer when prompted to reconsider, and (ii) quantitatively in terms of self-reported confidence score. We investigate the performance of three LLMs -- GPT4o, GPT4-turbo and Mistral -- on two benchmark sets of questions on causal judgement and formal fallacies and a set of probability and statistical puzzles and paradoxes. Although the LLMs show significantly better performance than random guessing, there is a wide variability in their tendency to change their initial answers. There is a positive correlation between qualitative confidence and accuracy, but the overall accuracy for the second answer is often worse than for the first answer. There is a strong tendency to overstate the self-reported confidence score. Confidence is only partially explained by the underlying token-level probability. The material effects of prompting on qualitative confidence and the strong tendency for overconfidence indicate that current LLMs do not have any internally coherent sense of confidence.