phi 2

Microsoft · microsoft/phi-2

ProviderMicrosoft
Familyphi
Typellm-code
Statusactive
Released2023-12-13

Reported benchmark scores

These scores carry one collection date for the whole card (2024-07) and a source list rather than a source per score, so they cannot be attributed individually. They are shown as reported and are not verified evidence. Benchmark pages carry the reviewed, dated evidence where it exists.
BenchmarkCatalogue standingScore
ARC-Challengeunassessed61.0
BIG-Bench Hardunassessed48.8
GPQA Diamondunassessed27.2
GSM8Kunassessed55.0
HellaSwagunassessed74.9
IFEvalunassessed27.4
MATH-500unassessed2.9
MMLU: Abstract Algebraunassessed29.0
MMLU: Anatomyunassessed44.4
MMLU: Astronomyunassessed58.6
MMLU: Business Ethicsunassessed56.0
MMLU: Clinical Knowledgeunassessed60.4
MMLU: College Biologyunassessed66.7
MMLU: College Chemistryunassessed40.0
MMLU: College Computer Scienceunassessed41.0
MMLU: College Mathematicsunassessed38.0
MMLU: College Medicineunassessed59.0
MMLU: College Physicsunassessed37.3
MMLU: Computer Securityunassessed74.0
MMLU: Conceptual Physicsunassessed52.3
MMLU: Econometricsunassessed38.6
MMLU: Electrical Engineeringunassessed55.2
MMLU: Elementary Mathematicsunassessed46.0
MMLU: Formal Logicunassessed35.7
MMLU: Global Factsunassessed38.0
MMLU: High School Biologyunassessed67.1
MMLU: High School Chemistryunassessed47.3
MMLU: High School Computer Scienceunassessed64.0
MMLU: High School European Historyunassessed64.8
MMLU: High School Geographyunassessed75.3
MMLU: High School Government and Politicsunassessed80.8
MMLU: High School Macroeconomicsunassessed58.2
MMLU: High School Mathematicsunassessed33.0
MMLU: High School Microeconomicsunassessed62.2
MMLU: High School Physicsunassessed38.4
MMLU: High School Psychologyunassessed79.3
MMLU: High School Statisticsunassessed47.2
MMLU: High School US Historyunassessed66.7
MMLU: High School World Historyunassessed74.3
MMLU: Human Agingunassessed65.9
MMLU: Human Sexualityunassessed70.2
MMLU: International Lawunassessed72.7
MMLU: Jurisprudenceunassessed73.1
MMLU: Logical Fallaciesunassessed73.0
MMLU: Machine Learningunassessed48.2
MMLU: Managementunassessed70.9
MMLU: Marketingunassessed82.5
MMLU: Medical Geneticsunassessed64.0
MMLU: Miscellaneousunassessed69.2
MMLU: Moral Disputesunassessed66.5
MMLU: Moral Scenariosunassessed29.9
MMLU: Nutritionunassessed62.1
MMLU: Philosophyunassessed62.1
MMLU: Prehistoryunassessed62.0
MMLU-Prounassessed26.3
MMLU: Professional Accountingunassessed44.0
MMLU: Professional Lawunassessed43.7
MMLU: Professional Medicineunassessed48.2
MMLU: Professional Psychologyunassessed56.4
MMLU: Public Relationsunassessed62.7
MMLU: Security Studiesunassessed71.8
MMLU: Sociologyunassessed80.1
MMLU: US Foreign Policyunassessed77.0
MMLU: Virologyunassessed47.0
MMLU: World Religionsunassessed69.6
MuSRunassessed41.0
TruthfulQAunassessed44.2
WinoGrandeunassessed73.5

Data

This card as JSON · Edit on GitHub