Calibration as a First-Class Criterion in LLM Evaluation
Abstract
Calibration of language models -- the alignment between expressed or implicit confidence and empirical correctness -- is a well-studied subfield within NLP. Methods to measure it already exist. The problem is adoption: outside this subfield, NLP research regularly introduces new models, datasets, and benchmarks without checking whether the model's confidence scores are meaningful. We argue that this adoption gap is a major obstacle to trustworthy LLM evaluation. Miscalibration causes problems in two distinct areas: at deployment, where overconfident mistakes cause real harm, and inside the research pipeline, where methods like LLM-as-a-judge, synthetic data generation, and active learning rely on calibrated confidence without verifying it. Standard calibration metrics only require two inputs per example: a confidence score and a correctness judgment. Most benchmarks in use today already provide both, meaning calibration can be reported immediately. For open-ended generation, however, defining these two inputs is still an open challenge. We argue that each NLP subfield should pair its main performance metric with a calibration score and call for treating calibration as an essential property of every model rather than a niche topic.
Community
We present a position paper arguing that calibration should be measured as a primary criterion in LLM evaluation. We discuss the consequences of neglecting it for both deployment and development, and outline directions for closing this gap.
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
- Pinocchio: Fast Uncertainty Estimates for Black-Box Language Models (2026)
- Claim-Level Confidence Calibration for Reliable Decision Making with Large Language Models (2026)
- DirEAG: Dirichlet Evidence Aggregation for Calibrating Verbalized Confidence in Mathematical Reasoning (2026)
- An Analysis of Training-Free Self-Reported Confidence in Language Models (2026)
- Can You Trust the Confidence? ConfBench for Vision-Language Models on Document Extraction (2026)
- From token probabilities to calibrated confidence: An empirical study of mathematical question answering (2026)
- Provable Limits and Certified Deferral for Verbalized Uncertainty in Small Language Models (2026)
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
Get this paper in your agent:
hf papers read 2609.26489 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash 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