HiBayES: A Hierarchical Bayesian Modeling Framework for AI Evaluation Statistics
This work addresses the need for systematic uncertainty quantification in AI evaluations, especially for costly and complex tests, but it is incremental as it builds on existing Bayesian and GLM methods.
The authors tackled the challenge of robustly estimating AI system capabilities and quantifying uncertainty in evaluations, particularly in low-data scenarios, by introducing HiBayES, a hierarchical Bayesian modeling framework that supports robust inferences and principled uncertainty quantification.
As Large Language Models (LLMs) and other AI systems evolve, robustly estimating their capabilities from inherently stochastic outputs while systematically quantifying uncertainty in these estimates becomes increasingly important. Further, advanced AI evaluations often have a nested hierarchical structure, exhibit high levels of complexity, and come with high costs in testing the most advanced AI systems. To address these challenges, we introduce HiBayES, a generalizable Hierarchical Bayesian modeling framework for AI Evaluation Statistics. HiBayES supports robust inferences in classical question-answer benchmarks and advanced agentic evaluations, particularly in low-data scenarios (e.g., < 20 data points per evaluation). Built on Generalized Linear Models (GLMs), Bayesian data analysis, and formal model comparison, HiBayES provides principled uncertainty quantification and robust parameter estimation. This paper offers a comprehensive introduction to HiBayES, including illustrative examples, comparisons to conventional statistical methods, and practical guidance for implementing multilevel Bayesian GLMs. Additionally, we provide a HiBayES software package [4] (Beta version) for out-of-the-box implementation.