Model

LLaMA 33B

Meta · open weights

ai.meta.com

Knowledge & Reasoning

Knowledge & Reasoning
BenchmarkScoreMetricHarnessSourceConfidenceLink
ARC-Challenge57.8accuracyN/AIndependentmediumarxiv.org
ARC-Easy80.0accuracyN/AIndependentmediumarxiv.org
BBH50.0averageN/AIndependentmediumarxiv.org
BBH39.8averageN/ASelf-reportedmediumarxiv.org
BoolQ86.1scoreN/AIndependentmediumcrfm.stanford.edu
BoolQ83.1accuracyN/AIndependentmediumarxiv.org
BoolQ83.1scoreN/ASelf-reportedmediumarxiv.org
HellaSwag82.8accuracyN/AIndependentmediumarxiv.org
HellaSwag82.8overall_accuracyN/ASelf-reportedmediumarxiv.org
LAMBADA77.2scoreN/AIndependentmediumarxiv.org
MMLU58.7exact_matchN/AIndependentmediumarxiv.org
MMLU57.8average_(%)N/AIndependentmediumarxiv.org
MMLU57.8exact_matchN/ASelf-reportedmediumarxiv.org
Natural Questions24.9exact_matchN/AIndependentmediumarxiv.org
OpenBookQA58.6accuracyN/ASelf-reportedmediumarxiv.org
OpenBookQA58.6accuracyN/AIndependentmediumarxiv.org
PIQA82.3accuracyN/AIndependentmediumarxiv.org
PIQA82.3scoreN/ASelf-reportedmediumarxiv.org
PIQA82.3scoreN/AIndependentmediumarxiv.org
siqa50.4accuracyN/AIndependentmediumarxiv.org
Trivia QA83.8exact_matchN/ASelf-reportedmediumarxiv.org
WinoGrande76.0accuracyN/AIndependentmediumarxiv.org
WinoGrande76.0accuracyN/ASelf-reportedmediumarxiv.org

Math

Math
BenchmarkScoreMetricHarnessSourceConfidenceLink
GSM8K44.1exact_matchN/AIndependentmediumarxiv.org
GSM8K35.6accuracyN/AIndependentmediumarxiv.org
GSM8K35.6exact_matchN/ASelf-reportedmediumarxiv.org
MATH7.1accuracyN/AIndependentmediumarxiv.org

Coding

Coding
BenchmarkScoreMetricHarnessSourceConfidenceLink
MBPP30.2accuracyN/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…