MMLU: High School Chemistry

MMLU subject subset: Standard high-school chemistry curriculum.

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
Subcategorychemistry
Page statusactive
Metricaccuracy
Directionhigher_is_better
Unit%
Dataset size203
Dataset licenceMIT
PublisherUC Berkeley (original); Center for AI Safety (current host)

What it measures

Standard high-school chemistry curriculum. Questions are four-option multiple-choice, drawn from the MMLU test set's "chemistry" subcategory within the benchmark's "STEM" top-level group, and are graded on the single correct labelled option.

Task format

Four-option multiple-choice questions, graded on the single correct labelled option; commonly evaluated 5-shot, consistent with the rest of MMLU.

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
Yi 1.5 34B 32K01.AI70.92024-07
Mixtral 8x22B Instruct v0.1Mistral AI69.52026-04
Yi 1.5 34B01.AI69.52026-04
Meta Llama 3 70BMeta69.02024-07
Meta Llama 3 70B InstructMeta69.02026-04
Meta Llama 3 70B InstructNous Research69.02026-04
Yi 1.5 34B Chat01.AI66.52026-04
Yi 1.5 34B Chat 16K01.AI66.52024-07
Yi 1.5 9B Chat01.AI66.52024-07
Yi 34B 200K01.AI65.02024-07
Mixtral 8x7B v0.1Mistral AI63.52026-04
Yi 1.5 9B01.AI63.52024-07
Yi 1.5 9B 32K01.AI63.52024-07
Mixtral 8x7B Instruct v0.1Mistral AI62.12026-04
Nous Hermes 2 Yi 34BNous Research62.12024-07
Yi 1.5 6B01.AI62.12024-07
Yi 34B Chat01.AI62.12024-07
Yi 1.5 9B Chat 16K01.AI61.12024-07
Phi 3 mini 128K instructMicrosoft60.12024-07
Phi 3 mini 4K instructMicrosoft60.12024-07
Nous Hermes 2 Mixtral 8x7B DPONous Research59.12024-07
Yi 9B01.AI58.12024-07
Meta Llama 3 8BMeta54.72024-07
Meta Llama 3 8BNous Research54.72024-07
Mistral 7B v0.3Mistral AI54.22024-07
mistral 7B v0.3 bnb 4bitUnsloth54.22024-07
Yi 1.5 6B Chat01.AI53.22024-07
Yi 6B01.AI53.22024-07
Yi 6B Chat01.AI53.22024-07
gemma 7B itGoogle DeepMind52.22024-07
Hermes 2 Theta Llama 3 8BNous Research51.72024-07
Nous Hermes 2 SOLAR 10.7BNous Research51.72024-07
Mistral 7B Instruct v0.2Mistral AI51.22024-07
Meta Llama 3 8B InstructMeta50.72024-07
Meta Llama 3 8B InstructNous Research50.72024-07
Qwen2 1.5B InstructAlibaba / Qwen Team49.32024-07
Hermes 2 Pro Llama 3 8BNous Research47.82024-07
phi 2Microsoft47.32024-07
chatglm2 6BZhipu AI45.32024-07
falcon 40BTII44.32024-07
Qwen2 0.5B InstructAlibaba / Qwen Team42.42024-07
gemma 2BGoogle DeepMind40.92024-07
deepseek llm 7B baseDeepSeek35.02024-07
deepseek llm 7B chatDeepSeek35.02024-07
deepseek coder 6.7B instructDeepSeek32.52024-07
deepseek coder 6.7B baseDeepSeek29.62024-07
gemma 2B itGoogle DeepMind29.12024-07
OLMo 1B hfAllen AI28.62024-07
deepseek coder 1.3B baseDeepSeek27.62024-07
deepseek coder 1.3B instructDeepSeek27.62024-07

Data

This page as JSON · Edit on GitHub