MMLU: High School Statistics

Accuracy on MMLU's high school statistics questions, one of 57 subject tests of academic and professional knowledge.

Also known as: high_school_statistics

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categoryknowledge
SubcategorySTEM
Page statusactive
Metricaccuracy
Directionhigher_is_better
Unit%
Dataset size216
Dataset licenceMIT
PublisherUC Berkeley

What it measures

Introductory descriptive and inferential statistics at AP level: probability, distributions, hypothesis testing, confidence intervals, and basic regression. Framed as four-option multiple-choice questions and scored zero-shot or few-shot by exact match against the labelled option, as one of the 57 subject subsets that make up the MMLU benchmark.

Task format

Four-option multiple-choice question answering (A-D), one correct answer, zero-shot or few-shot.

Models reporting this benchmark

These figures come from the model cards, which carry one collection date per card and no per-score attribution. They are shown as reported, not as verified evidence.
ModelProviderScoreCard as of
Meta Llama 3 70BMeta73.62024-07
Meta Llama 3 70B InstructMeta73.62026-04
Meta Llama 3 70B InstructNous Research73.62026-04
Yi 1.5 34B01.AI72.72026-04
Yi 1.5 34B 32K01.AI71.32024-07
Mixtral 8x22B Instruct v0.1Mistral AI69.42026-04
Yi 1.5 9B Chat01.AI68.52024-07
Yi 9B01.AI68.52024-07
Yi 1.5 34B Chat01.AI68.12026-04
Yi 1.5 34B Chat 16K01.AI68.12024-07
Yi 34B 200K01.AI67.12024-07
Nous Hermes 2 Mixtral 8x7B DPONous Research66.22024-07
Nous Hermes 2 Yi 34BNous Research66.22024-07
Yi 1.5 9B 32K01.AI66.22024-07
Yi 1.5 9B01.AI65.32024-07
Yi 1.5 9B Chat 16K01.AI65.32024-07
Mixtral 8x7B v0.1Mistral AI64.82026-04
Phi 3 mini 128K instructMicrosoft64.82024-07
Yi 34B Chat01.AI63.42024-07
Phi 3 mini 4K instructMicrosoft61.62024-07
Mixtral 8x7B Instruct v0.1Mistral AI59.72026-04
gemma 7B itGoogle DeepMind57.92024-07
Meta Llama 3 8BMeta57.92024-07
Meta Llama 3 8BNous Research57.92024-07
Yi 1.5 6B01.AI57.42024-07
Meta Llama 3 8B InstructMeta53.72024-07
Meta Llama 3 8B InstructNous Research53.72024-07
Yi 1.5 6B Chat01.AI53.22024-07
Nous Hermes 2 SOLAR 10.7BNous Research52.82024-07
Mistral 7B v0.3Mistral AI51.92024-07
mistral 7B v0.3 bnb 4bitUnsloth51.92024-07
Yi 6B01.AI51.92024-07
Yi 6B Chat01.AI51.92024-07
Hermes 2 Pro Llama 3 8BNous Research50.92024-07
Hermes 2 Theta Llama 3 8BNous Research50.92024-07
falcon 40BTII47.72024-07
phi 2Microsoft47.22024-07
Mistral 7B Instruct v0.2Mistral AI44.42024-07
Qwen2 1.5B InstructAlibaba / Qwen Team44.42024-07
deepseek llm 7B baseDeepSeek43.52024-07
deepseek llm 7B chatDeepSeek43.52024-07
Qwen2 0.5B InstructAlibaba / Qwen Team38.92024-07
deepseek coder 6.7B baseDeepSeek37.02024-07
deepseek coder 6.7B instructDeepSeek36.62024-07
OLMo 1B hfAllen AI36.62024-07
gemma 2BGoogle DeepMind35.22024-07
chatglm2 6BZhipu AI33.82024-07
deepseek coder 1.3B baseDeepSeek25.52024-07
deepseek coder 1.3B instructDeepSeek25.52024-07
gemma 2B itGoogle DeepMind20.42024-07

Data

This page as JSON · Edit on GitHub