Confidence Diagram of Nonparametric Ranking for Uncertainty Assessment in Large Language Models Evaluation

2024-12-07

AI Summary

This work addresses two key challenges in large language model (LLM) evaluation: high uncertainty in model ranking and the difficulty of reliably characterizing domain-specific capabilities. To this end, we propose a nonparametric contextual ranking framework. Methodologically, we introduce the novel concept of a “confidence graph”—a Hasse diagram that uniformly encodes all statistically significant, confidence-supported orderings; extend the Gaussian multiplier bootstrap to non-i.i.d. empirical processes to enable rigorous upper-bound inference for rankings; and integrate context-aware scoring, compositional reasoning, and multiple hypothesis testing. Experiments on synthetic and real-world medical datasets demonstrate that our framework substantially improves the reliability and interpretability of LLM rankings for domain-specific competencies—particularly clinical reasoning—and provides the first theoretically grounded uncertainty quantification tool for Best-of-N alignment strategies.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Reasoning under Uncertainty: Graphical Models
• Natural Language Processing: (Large) Language Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

We consider the inference for the ranking of large language models (LLMs). Alignment arises as a significant challenge to mitigate hallucinations in the use of LLMs. Ranking LLMs has proven to be an effective tool to improve alignment based on the best-of-$N$ policy. In this paper, we propose a new inferential framework for hypothesis testing among the ranking for language models. Our framework is based on a nonparametric contextual ranking framework designed to assess large language models' domain-specific expertise, leveraging nonparametric scoring methods to account for their sensitivity to the prompts. To characterize the combinatorial complexity of the ranking, we introduce a novel concept of confidence diagram, which leverages a Hasse diagram to represent the entire confidence set of rankings by a single directed graph. We show the validity of the proposed confidence diagram by advancing the Gaussian multiplier bootstrap theory to accommodate the supremum of independent empirical processes that are not necessarily identically distributed. Extensive numerical experiments conducted on both synthetic and real data demonstrate that our approach offers valuable insight into the evaluation for the performance of different LLMs across various medical domains.

Problem

• Inference for ranking large language models
• Mitigate hallucinations using nonparametric ranking
• Evaluate LLMs' domain-specific expertise

Innovation

• Nonparametric contextual ranking framework
• Confidence diagram using Hasse diagram
• Gaussian multiplier bootstrap theory