Confident Rankings with Fewer Items: Adaptive LLM Evaluation with Continuous Scores

arXiv:2601.13885v2 Announce Type: replace-cross Abstract: Computerized Adaptive Testing (CAT) has proven effective for efficient LLM evaluation on multiple-choice benchmarks, but modern LLM evaluation increasingly relies on generation tasks where outputs are scored continuously rather than marked correct/incorrect. We present a principled extension of IRT-based adaptive testing to continuous bounded scores (ROUGE, BLEU, LLM-as-a-Judge) by replacing the Bernoulli response distribution with a heteroskedastic normal distribution. Building on this, we introduce an uncertainty aware ranker with adaptive stopping criteria that achieves reliable model ranking while testing as few items and as cheaply as possible. We validate our method on five benchmarks spanning n-gram-based, embedding-based, and LLM-as-judge metrics. Our method improves ranking correlation by 0.13 au over random sampling and has 99% accuracy on confident predictions while using 2% of the items after a one-time calibration step.

Related

• Inclusion-of-Thoughts: Mitigating Preference Instability via Purifying the Decision Space
• RankLLM: Weighted Ranking of LLMs by Quantifying Question Difficulty
• Response-free item difficulty modelling for multiple-choice items with fine-tuned transformers: Component-wise representation and multi-task learning

Source: arXiv cs.AI | 2026-09-15