# Language Models Prefer What They Know: Relative Confidence Estimation via Confidence Preferences

arXiv: https://arxiv.org/abs/2502.01126 · HTML: https://mlsanity.com/paper/2502.01126/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 Feb 2025.

## Built on by (most cited first, 13 of 13)

- [Uncertainty Quantification and Confidence Calibration in Large Language Models: A Survey](https://mlsanity.com/paper/2503.15850/) (2025)
- [No Answer Needed: Predicting LLM Answer Accuracy from Question-Only Linear Probes](https://mlsanity.com/paper/2509.10625/) (2025)
- [Always Tell Me The Odds: Fine-grained Conditional Probability Estimation](https://mlsanity.com/paper/2505.01595/) (2025)
- [Linguistic Nepotism: Trading-off Quality for Language Preference in Multilingual RAG](https://mlsanity.com/paper/2509.13930/) (2025)
- [Self-Evaluating LLMs for Multi-Step Tasks: Stepwise Confidence Estimation for Failure Detection](https://mlsanity.com/paper/2511.07364/) (2025)
- [Generalization of Fine-Tuned Uncertainty Communication and Metacognition in Large Language Models](https://mlsanity.com/paper/2510.05126/) (2025)
- [Confidence Before Answering: A Paradigm Shift for Efficient LLM Uncertainty Estimation](https://mlsanity.com/paper/2603.05881/) (2026)
- [Confident RAG: Enhancing the Performance of LLMs for Mathematics Question Answering through Multi-Embedding and Confidence Scoring](https://mlsanity.com/paper/2507.17442/) (2025)
- [Beware of Reasoning Overconfidence: Pitfalls in the Reasoning Process for Multi-solution Tasks](https://mlsanity.com/paper/2512.01725/) (2025)
- [Quantifying Consistency in LLM Logical Reasoning via Structural Uncertainty](https://mlsanity.com/paper/2606.17312/) (2026)
- [Reasoning Strategies in Large Language Models: Can They Follow, Prefer, and Optimize?](https://mlsanity.com/paper/2507.11423/) (2025)
- [CLSGen: A Dual-Head Fine-Tuning Framework for Joint Probabilistic Classification and Verbalized Explanation](https://mlsanity.com/paper/2604.11801/) (2026)
- [READY or Not: Reliable Enterprise Agent Deployment](https://mlsanity.com/paper/2609.02095/) (2026)

## Abstract (arXiv)

Language models (LMs) should provide reliable confidence estimates to help users detect mistakes in their outputs and defer to human experts when necessary. Asking a language model to assess its confidence ("Score your confidence from 0-1.") is a natural way of evaluating its uncertainty. However, models struggle to provide absolute assessments of confidence (i.e. judging confidence in answering a question independent of other questions) and the coarse-grained scores they produce are not useful for evaluating the correctness of their answers. We propose relative confidence estimation, where we match up questions against each other and ask the model to make relative judgments of confidence ("Which question are you more confident in answering correctly?"). Treating each question as a "player" in a series of matchups against other questions and the model's preferences as match outcomes, we can use rank aggregation methods like Elo rating and Bradley-Terry to translate the model's confidence preferences into confidence scores. We evaluate relative confidence estimation against absolute confidence estimation and self-consistency confidence methods on five state-of-the-art LMs -- GPT-4, GPT-4o, Gemini 1.5 Pro, Claude 3.5 Sonnet, and Llama 3.1 405B -- across 14 challenging STEM, social science, and commonsense reasoning question answering tasks. Our results demonstrate that relative confidence estimation consistently provides more reliable confidence scores than absolute confidence estimation, with average gains of 3.5% in selective classification AUC over direct absolute confidence estimation methods and 1.7% over self-consistency approaches across all models and datasets.