MMLU: High School Psychology

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

Also known as: high_school_psychology

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
SubcategorySocial Sciences
Page statusactive
Metricaccuracy
Directionhigher_is_better
Unit%
Dataset size545
Dataset licenceMIT
PublisherUC Berkeley

What it measures

Introductory psychology at AP level: cognition, learning theory, developmental and social psychology, personality, sensation and perception, and research methods. 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 70BMeta94.12024-07
Meta Llama 3 70B InstructMeta94.12026-04
Meta Llama 3 70B InstructNous Research94.12026-04
Mixtral 8x22B Instruct v0.1Mistral AI92.72026-04
Nous Hermes 2 Yi 34BNous Research92.12024-07
Yi 34B 200K01.AI92.12024-07
Yi 1.5 34B01.AI91.42026-04
Yi 1.5 34B 32K01.AI91.02024-07
Yi 34B Chat01.AI90.82024-07
Nous Hermes 2 Mixtral 8x7B DPONous Research89.52024-07
Phi 3 mini 128K instructMicrosoft89.02024-07
Yi 1.5 34B Chat01.AI89.02026-04
Yi 1.5 34B Chat 16K01.AI89.02024-07
Mixtral 8x7B Instruct v0.1Mistral AI88.42026-04
Phi 3 mini 4K instructMicrosoft88.32024-07
Mixtral 8x7B v0.1Mistral AI88.12026-04
Yi 9B01.AI86.82024-07
Yi 1.5 9B 32K01.AI86.42024-07
Yi 1.5 9B Chat01.AI86.42024-07
Yi 1.5 9B01.AI85.92024-07
Yi 1.5 9B Chat 16K01.AI85.72024-07
Nous Hermes 2 SOLAR 10.7BNous Research85.12024-07
Meta Llama 3 8BMeta84.82024-07
Meta Llama 3 8BNous Research84.82024-07
Meta Llama 3 8B InstructMeta83.72024-07
Meta Llama 3 8B InstructNous Research83.72024-07
gemma 7B itGoogle DeepMind83.12024-07
Hermes 2 Theta Llama 3 8BNous Research83.12024-07
Yi 1.5 6B01.AI83.12024-07
Hermes 2 Pro Llama 3 8BNous Research82.92024-07
Yi 6B01.AI82.82024-07
Yi 6B Chat01.AI82.82024-07
Yi 1.5 6B Chat01.AI82.42024-07
Mistral 7B v0.3Mistral AI80.02024-07
mistral 7B v0.3 bnb 4bitUnsloth80.02024-07
phi 2Microsoft79.32024-07
Mistral 7B Instruct v0.2Mistral AI79.12024-07
falcon 40BTII77.62024-07
Qwen2 1.5B InstructAlibaba / Qwen Team75.22024-07
deepseek llm 7B baseDeepSeek71.22024-07
deepseek llm 7B chatDeepSeek71.22024-07
chatglm2 6BZhipu AI59.32024-07
gemma 2BGoogle DeepMind57.22024-07
Qwen2 0.5B InstructAlibaba / Qwen Team56.52024-07
gemma 2B itGoogle DeepMind51.02024-07
deepseek coder 6.7B instructDeepSeek41.72024-07
deepseek coder 6.7B baseDeepSeek38.52024-07
deepseek coder 1.3B baseDeepSeek31.72024-07
deepseek coder 1.3B instructDeepSeek31.72024-07
OLMo 1B hfAllen AI24.02024-07

Data

This page as JSON · Edit on GitHub