MMLU: Moral Disputes

Accuracy on MMLU's moral disputes questions, one of 57 subject tests of academic and professional knowledge.

Also known as: moral_disputes

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
SubcategoryHumanities
Page statusactive
Metricaccuracy
Directionhigher_is_better
Unit%
Dataset size346
Dataset licenceMIT
PublisherUC Berkeley

What it measures

Applied-ethics questions covering disputed moral issues, testing whether a model can identify and reason about the philosophical positions at stake rather than settle the disputes itself. 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
Mixtral 8x22B Instruct v0.1Mistral AI84.72026-04
Meta Llama 3 70BMeta84.42024-07
Meta Llama 3 70B InstructMeta84.42026-04
Meta Llama 3 70B InstructNous Research84.42026-04
Nous Hermes 2 Yi 34BNous Research83.52024-07
Yi 1.5 34B 32K01.AI82.12024-07
Yi 1.5 34B01.AI81.22026-04
Yi 1.5 34B Chat01.AI81.22026-04
Yi 1.5 34B Chat 16K01.AI81.22024-07
Yi 34B Chat01.AI80.92024-07
Yi 34B 200K01.AI80.32024-07
Nous Hermes 2 Mixtral 8x7B DPONous Research80.12024-07
Mixtral 8x7B v0.1Mistral AI79.82026-04
Mixtral 8x7B Instruct v0.1Mistral AI78.02026-04
Yi 9B01.AI76.62024-07
Yi 1.5 9B 32K01.AI75.72024-07
Phi 3 mini 4K instructMicrosoft75.42024-07
Hermes 2 Theta Llama 3 8BNous Research74.92024-07
Meta Llama 3 8B InstructMeta74.92024-07
Meta Llama 3 8B InstructNous Research74.92024-07
Yi 1.5 9B01.AI74.62024-07
Meta Llama 3 8BMeta74.02024-07
Meta Llama 3 8BNous Research74.02024-07
Nous Hermes 2 SOLAR 10.7BNous Research74.02024-07
Phi 3 mini 128K instructMicrosoft74.02024-07
Yi 1.5 9B Chat01.AI73.72024-07
Yi 1.5 9B Chat 16K01.AI72.32024-07
gemma 7B itGoogle DeepMind71.72024-07
Hermes 2 Pro Llama 3 8BNous Research71.12024-07
Mistral 7B v0.3Mistral AI70.22024-07
mistral 7B v0.3 bnb 4bitUnsloth70.22024-07
Yi 6B01.AI69.92024-07
Yi 6B Chat01.AI69.92024-07
Mistral 7B Instruct v0.2Mistral AI69.72024-07
Yi 1.5 6B Chat01.AI69.72024-07
Yi 1.5 6B01.AI69.42024-07
phi 2Microsoft66.52024-07
falcon 40BTII62.72024-07
Qwen2 1.5B InstructAlibaba / Qwen Team62.72024-07
deepseek llm 7B baseDeepSeek59.02024-07
deepseek llm 7B chatDeepSeek59.02024-07
chatglm2 6BZhipu AI52.92024-07
Qwen2 0.5B InstructAlibaba / Qwen Team51.22024-07
gemma 2BGoogle DeepMind43.92024-07
deepseek coder 6.7B instructDeepSeek41.92024-07
gemma 2B itGoogle DeepMind40.82024-07
deepseek coder 6.7B baseDeepSeek40.52024-07
deepseek coder 1.3B baseDeepSeek28.02024-07
deepseek coder 1.3B instructDeepSeek28.02024-07
OLMo 1B hfAllen AI25.12024-07

Data

This page as JSON · Edit on GitHub