MMLU-Pro Leaderboard

Introduction

We introduce MMLU-Pro, an enhanced benchmark designed to evaluate language understanding models across broader and more challenging tasks. Building on the Massive Multitask Language Understanding (MMLU) dataset, MMLU-Pro integrates more challenging, reasoning-focused questions and increases the answer choices per question from four to ten, significantly raising the difficulty and reducing the chance of success through random guessing. MMLU-Pro comprises over 12,000 rigorously curated questions from academic exams and textbooks, spanning 14 diverse domains including Biology, Business, Chemistry, Computer Science, Economics, Engineering, Health, History, Law, Math, Philosophy, Physics, Psychology, and Others.

Note: For inclusion in our leaderboard, submissions must provide substantial evidence demonstrating that their system is a genuine language model. We maintain strict verification standards to ensure the integrity and comparability of the results.

What's new about MMLU-Pro

Compared to the original MMLU, there are three major differences:

• The original MMLU dataset only contains 4 options, MMLU-Pro increases it to 10 options. The increase in options will make the evaluation more realistic and challenging. The random guessing will lead to a much lower score.
• The original MMLU dataset contains mostly knowledge-driven questions without requiring much reasoning. Therefore, PPL results are normally better than CoT. In our dataset, we increase the problem difficulty and integrate more reasoning-focused problems. In MMLU-Pro, CoT can be 20% higher than PPL.
• By increasing the distractor numbers, we significantly reduce the probability of correct guess by chance to boost the benchmark’s robustness. Specifically, with 24 different prompt styles tested, the sensitivity of model scores to prompt variations decreased from 4-5% in MMLU to just 2% in MMLU-Pro.

For detailed information about the dataset, visit our page on Hugging Face: https://huggingface.co/datasets/TIGER-Lab/MMLU-Pro.

If you are interested in replicating these results or wish to evaluate your models using our dataset, access our evaluation scripts available on GitHub: https://github.com/TIGER-AI-Lab/MMLU-Pro.

If you would like to learn more details about our dataset, please check out our paper: https://arxiv.org/abs/2406.01574.

Below you can find the accuracies of different models tested on this dataset.