Are We Done with MMLU?
Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:GDS7MHYHrecord.jsonopen to challenge →
read the original abstract
Maybe not. We identify and analyse errors in the popular Massive Multitask Language Understanding (MMLU) benchmark. Even though MMLU is widely adopted, our analysis demonstrates numerous ground truth errors that obscure the true capabilities of LLMs. For example, we find that 57% of the analysed questions in the Virology subset contain errors. To address this issue, we introduce a comprehensive framework for identifying dataset errors using a novel error annotation protocol. Then, we create MMLU-Redux, which is a subset of 5,700 manually re-annotated questions across all 57 MMLU subjects. We estimate that 6.49% of MMLU questions contain errors. Using MMLU-Redux, we demonstrate significant discrepancies with the model performance metrics that were originally reported. Our results strongly advocate for revising MMLU's error-ridden questions to enhance its future utility and reliability as a benchmark. https://huggingface.co/datasets/edinburgh-dawg/mmlu-redux-2.0.
This paper has not been read by Pith yet.
Forward citations
Cited by 23 Pith papers
-
Zone of Proximal Policy Optimization: Teacher in Prompts, Not Gradients
ZPPO improves distillation to small vision-language models by using binary and negative candidate prompts plus a replay buffer for hard questions, outperforming standard distillation and GRPO on a 31-benchmark suite w...
-
Knowledge Index of Noah's Ark
Introduces KINA benchmark with 899 items over 261 disciplines, formal (1-1/e) coverage guarantee and bonus-on-bar tournament theorem, plus evaluations of 42 models with top score 53.17%.
-
Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing
Magpie synthesizes 300K high-quality alignment instructions from Llama-3-Instruct via auto-regressive prompting on partial templates, enabling fine-tuned models to match official instruct performance on AlpacaEval, Ar...
-
Improved Large Language Diffusion Models
iLLaDA is an 8B masked diffusion LM trained from scratch with bidirectional attention, reporting gains of 14-21 points on BBH, ARC, MATH and HumanEval over prior diffusion models while remaining competitive with Qwen2.5-7B.
-
Dynamic Model Merging Made Slim
DiDi-Merging achieves dynamic model merging performance matching or exceeding prior methods while using only 1.24x to 1.4x the parameters of a single fine-tuned model.
-
SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training
Pruning pretrained MoE models outperforms training from scratch, different compression methods converge after continued pretraining, and combining KD with language modeling loss plus progressive schedules yields a com...
-
Kimi Linear: An Expressive, Efficient Attention Architecture
Kimi Linear hybridizes linear attention with a new KDA module to beat full attention on tasks while slashing KV cache by 75% and speeding decoding up to 6x.
-
Qwen3-Omni Technical Report
Qwen3-Omni is a unified multimodal model that achieves open-source SOTA on 32 of 36 audio and audio-visual benchmarks and overall SOTA on 22 without degrading performance on text, image, or video relative to single-mo...
-
Mixture-of-Experts Can Surpass Dense LLMs Under Strictly Equal Resource
MoE models with activation rates in an optimal region outperform dense LLMs of identical total parameter count, training compute, and data budget, with the optimal region consistent across scales.
-
Qwen2.5-1M Technical Report
Qwen2.5-1M models reach 1M token context with improved long-context performance, no short-context loss, and 3-7x prefill speedup via open inference optimizations.
-
Yuvion LLM: An Adversarially-Aware Large Language Model for Content And AI Safety
Yuvion LLM applies adversarially aware training and introduces the YLRE benchmark set, claiming superior safety robustness over larger models on multiple tasks.
-
SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training
Pruning pretrained MoE models outperforms training from scratch under fixed budget, different expert compression methods converge after continued training, and progressive pruning plus multi-token KD improves the fina...
-
Qwen3.5-Omni Technical Report
Qwen3.5-Omni scales an omnimodal model to hundreds of billions of parameters with 256k context, introduces ARIA for stable speech synthesis, and reports SOTA performance on 215 audio-visual benchmarks while adding mul...
-
SnapMLA: Efficient Long-Context MLA Decoding via Hardware-Aware FP8 Quantized Pipelining
SnapMLA achieves up to 1.91x higher throughput in long-output MLA decoding using FP8 quantization and specialized kernels while keeping benchmark quality near the BF16 baseline.
-
MiMo-V2-Flash Technical Report
MiMo-V2-Flash is a 309B/15B MoE model trained on 27T tokens with hybrid attention and multi-teacher on-policy distillation that matches larger models like DeepSeek-V3.2 while enabling 2.6x faster decoding via repurpos...
-
Kimi K2: Open Agentic Intelligence
Kimi K2 is a 1-trillion-parameter MoE model that leads open-source non-thinking models on agentic benchmarks including 65.8 on SWE-Bench Verified and 66.1 on Tau2-Bench.
-
Qwen3 Technical Report
Pith review generated a malformed one-line summary.
-
Qwen2.5-Omni Technical Report
Qwen2.5-Omni presents a multimodal model with block-wise encoders, TMRoPE position embeddings, and a Thinker-Talker architecture that enables simultaneous text and streaming speech generation while matching text perfo...
-
Qwen2.5-VL Technical Report
Qwen2.5-VL reports a vision-language model family using native dynamic-resolution ViT and absolute time encoding that matches GPT-4o on document and diagram tasks while supporting hour-long videos with second-level lo...
-
Measuring AI Reasoning: A Guide for Researchers
Reasoning in language models should be measured by the faithfulness and validity of their multi-step search processes and intermediate traces, not final-answer accuracy.
-
Ministral 3
Ministral 3 releases 3B/8B/14B parameter-efficient language models with base, instruction, and reasoning variants derived via iterative pruning and distillation, including image understanding capabilities.
-
Position: Stop Evaluating AI with Human Tests, Develop Principled, AI-specific Tests instead
Human tests should not be applied to AI to measure traits like intelligence due to calibration, validity, contamination, and prompt sensitivity issues; develop AI-specific evaluation frameworks instead.
-
Mellum2 Technical Report
Mellum 2 is a 12B MoE model with 2.5B active parameters, trained on 10.6T tokens with MoE, GQA, SWA, and MTP, then post-trained into Instruct and Thinking variants, claimed competitive with 4B-14B models at 2.5B compute.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.