Judging with Confidence: Calibrating Autoraters to Preference Distributions
Abstract
A framework for calibrating probabilistic autoraters to preference distributions using supervised fine-tuning and reinforcement learning improves alignment with human values and reduces bias.
The alignment of large language models (LLMs) with human values increasingly relies on using other LLMs as automated judges, or ``autoraters''. However, their reliability is limited by a foundational issue: they are trained on discrete preference labels, forcing a single ground truth onto tasks that are often subjective, ambiguous, or nuanced. We argue that a reliable autorater must learn to model the full distribution of preferences defined by a target population. In this paper, we propose a general framework for calibrating probabilistic autoraters to any given preference distribution. We formalize the problem and present two learning methods tailored to different data conditions: 1) a direct supervised fine-tuning for dense, probabilistic labels, and 2) a reinforcement learning approach for sparse, binary labels. Our empirical results show that finetuning autoraters with a distribution-matching objective leads to verbalized probability predictions that are better aligned with the target preference distribution, with improved calibration and significantly lower positional bias, all while preserving performance on objective tasks.
Community
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- MENLO: From Preferences to Proficiency -- Evaluating and Modeling Native-like Quality Across 47 Languages (2025)
- Can Large Language Models Express Uncertainty Like Human? (2025)
- S2J: Bridging the Gap Between Solving and Judging Ability in Generative Reward Models (2025)
- Latent Collective Preference Optimization: A General Framework for Robust LLM Alignment (2025)
- ConfTuner: Training Large Language Models to Express Their Confidence Verbally (2025)
- From Scores to Preferences: Redefining MOS Benchmarking for Speech Quality Reward Modeling (2025)
- Reference-Free Rating of LLM Responses via Latent Information (2025)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment:
@librarian-bot
recommend
Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper