PredictaBoard: Benchmarking LLM Score Predictability

TL;DR

PredictaBoard introduces a benchmarking framework to evaluate how well assessors can predict LLM errors, aiming to improve the safety and reliability of large language models by focusing on their predictability.

Contribution

This paper presents a novel collaborative benchmarking framework, PredictaBoard, for evaluating the ability of score predictors to anticipate LLM errors on specific tasks.

Findings

Baseline assessors show varying prediction accuracy

PredictaBoard reveals the importance of predictability in LLM safety

Framework encourages development of more reliable LLM assessors

Abstract

Despite possessing impressive skills, Large Language Models (LLMs) often fail unpredictably, demonstrating inconsistent success in even basic common sense reasoning tasks. This unpredictability poses a significant challenge to ensuring their safe deployment, as identifying and operating within a reliable "safe zone" is essential for mitigating risks. To address this, we present PredictaBoard, a novel collaborative benchmarking framework designed to evaluate the ability of score predictors (referred to as assessors) to anticipate LLM errors on specific task instances (i.e., prompts) from existing datasets. PredictaBoard evaluates pairs of LLMs and assessors by considering the rejection rate at different tolerance errors. As such, PredictaBoard stimulates research into developing better assessors and making LLMs more predictable, not only with a higher average performance. We conduct illustrative experiments using baseline assessors and state-of-the-art LLMs. PredictaBoard highlights the critical need to evaluate predictability alongside performance, paving the way for safer AI systems where errors are not only minimised but also anticipated and effectively mitigated. Code for our benchmark can be found at https://github.com/Kinds-of-Intelligence-CFI/PredictaBoard