MR-Ben: A Meta-Reasoning Benchmark for Evaluating System-2 Thinking in LLMs

TL;DR

MR-Ben is a new benchmark designed to evaluate large language models' ability to perform meta-reasoning, specifically in identifying errors in their reasoning steps, which is crucial for system-2 slow thinking.

Contribution

The paper introduces MR-Ben, a comprehensive process-based benchmark for assessing meta-reasoning skills in LLMs, highlighting current model limitations.

Findings

Open-source models show weaker meta-reasoning abilities.

Models like OpenAI's o1 series excel at scrutinizing solutions.

Benchmark reveals weaknesses in existing LLM training and inference methods.

Abstract

Large language models (LLMs) have shown increasing capability in problem-solving and decision-making, largely based on the step-by-step chain-of-thought reasoning processes. However, evaluating these reasoning abilities has become increasingly challenging. Existing outcome-based benchmarks are beginning to saturate, becoming less effective in tracking meaningful progress. To address this, we present a process-based benchmark MR-Ben that demands a meta-reasoning skill, where LMs are asked to locate and analyse potential errors in automatically generated reasoning steps. Our meta-reasoning paradigm is especially suited for system-2 slow thinking, mirroring the human cognitive process of carefully examining assumptions, conditions, calculations, and logic to identify mistakes.MR-Ben comprises 5,975 questions curated by human experts across a wide range of subjects, including physics, chemistry, logic, coding, and more. Through our designed metrics for assessing meta-reasoning on this benchmark, we identify interesting limitations and weaknesses of current LLMs (open-source and closed-source models). For example, with models like the o1 series from OpenAI demonstrating strong performance by effectively scrutinizing the solution space, many other state-of-the-art models fall significantly behind on MR-Ben, exposing potential shortcomings in their training strategies and inference methodologies.