Multiple-Choice Questions are Efficient and Robust LLM Evaluators

TL;DR

This paper introduces GSM-MC, a multiple-choice dataset for evaluating large language models, demonstrating its efficiency, robustness, and strong correlation with original benchmarks, while also presenting new MC datasets for math and programming tasks.

Contribution

The paper creates and validates multiple-choice evaluation datasets for LLMs, showing their effectiveness and introducing new benchmarks for math and programming reasoning.

Findings

MC evaluation correlates strongly with original performance

MC evaluation reduces assessment time significantly

LLMs still have substantial room for improvement on these benchmarks

Abstract

We present GSM-MC, a multiple-choice (MC) dataset constructed by collecting answers and incorrect predictions on GSM8K from 60 open-source models. Through extensive experiments, we show that LLMs' performance on the MC version of this popular benchmark is strongly correlated with their performance on the original version and is quite robust to distractor choices and option orders, while the evaluation time is reduced by a factor of up to 30. Following similar procedures, we introduce MATH-MC, constructed from MATH, and PythonIO, a new program reasoning MC dataset constructed from HumanEval and MBPP. Experimental results indicate that LLMs' performance on these MC benchmarks leaves much room for improvement. Our data and code are available at https://github.com/Geralt-Targaryen/MC-Evaluation.