CriticAL: Critic Automation with Language Models

TL;DR

CriticAL leverages large language models to automate scientific model criticism by generating and evaluating discrepancies between models and data, improving model validation and development.

Contribution

This paper introduces CriticAL, a novel framework that automates model criticism using LLMs within a hypothesis testing approach, addressing hallucination issues and enhancing scientific discovery.

Findings

CriticAL reliably generates accurate critiques without hallucinations.

CriticAL's critiques are preferred for transparency and actionability.

CriticAL enables LLM scientists to improve models on real datasets.

Abstract

Understanding the world through models is a fundamental goal of scientific research. While large language model (LLM) based approaches show promise in automating scientific discovery, they often overlook the importance of criticizing scientific models. Criticizing models deepens scientific understanding and drives the development of more accurate models. Automating model criticism is difficult because it traditionally requires a human expert to define how to compare a model with data and evaluate if the discrepancies are significant--both rely heavily on understanding the modeling assumptions and domain. Although LLM-based critic approaches are appealing, they introduce new challenges: LLMs might hallucinate the critiques themselves. Motivated by this, we introduce CriticAL (Critic Automation with Language Models). CriticAL uses LLMs to generate summary statistics that capture discrepancies between model predictions and data, and applies hypothesis tests to evaluate their significance. We can view CriticAL as a verifier that validates models and their critiques by embedding them in a hypothesis testing framework. In experiments, we evaluate CriticAL across key quantitative and qualitative dimensions. In settings where we synthesize discrepancies between models and datasets, CriticAL reliably generates correct critiques without hallucinating incorrect ones. We show that both human and LLM judges consistently prefer CriticAL's critiques over alternative approaches in terms of transparency and actionability. Finally, we show that CriticAL's critiques enable an LLM scientist to improve upon human-designed models on real-world datasets.