On the Stability of Prompt Ranking in Large Language Model Evaluation

2026-06-23

AI Summary

This work addresses the high sensitivity of prompt ranking in large language models to evaluation perturbations—such as variations in random seeds or test subsets—which undermines the reliability of selecting optimal prompts. The study systematically analyzes the stability of prompt rankings under diverse evaluation conditions and, for the first time, reveals the fragility of top-ranked prompts. To mitigate this issue, the authors propose a stability-aware selection strategy based on lower confidence bounds. By integrating multi-model, multi-task experiments with statistical confidence interval estimation, the method balances performance and variance, significantly enhancing selection robustness in unstable scenarios while maintaining competitive performance in stable ones. These findings underscore the critical importance of modeling evaluation uncertainty for reliable prompt selection.

Abstract

Prompt-based interaction has become a dominant paradigm for using large language models (LLMs), where multiple candidate prompts are evaluated and the top-ranked one is selected for downstream use. This workflow implicitly assumes that prompt rankings are stable under minor variations in evaluation conditions. In this paper, we systematically study prompt ranking stability under common sources of variability, including random seeds and limited evaluation subsets. Across three open-weight LLMs and two benchmark tasks, we find that while overall rank correlations are often moderate to high, the identity of the top-performing prompt frequently changes, leading to unreliable selection decisions. To address this issue, we propose a simple stability-aware selection strategy based on a lower confidence bound, which accounts for both performance and variance. Our results show that this approach improves robustness in unstable settings while remaining competitive in more stable regimes. These findings highlight the importance of accounting for evaluation uncertainty in prompt selection and LLM benchmarking.

Problem

• prompt ranking
• stability
• large language models
• evaluation uncertainty
• benchmarking

Innovation

• prompt ranking stability
• large language models
• evaluation uncertainty
• confidence-bound selection
• robust prompt selection