Quantifying Ranking Uncertainty in LLM Benchmarks
Bitya Neuhof, Yuval Benjamini
cs.LG, cs.CL, stat.ML
Submitted: 2026-06-28
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Statistical multi-metric evaluation and visualization of LLM system predictive performance
- Statistical Uncertainty Quantification for Aggregate Performance Metrics in Machine Learning Benchmarks
- Adding Error Bars to Evals: A Statistical Approach to Language Model Evaluations
- Finite-Sample Valid Rank Confidence Sets for a Broad Class of Statistical and Machine Learning Models
- csranks: An R Package for Estimation and Inference Involving Ranks
- Uncertainty in Ranking
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks