Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Question Answering on Commonsense QA (ARC, HellaSwag, OBQA, RTE, CoPa, Race)

82.83ARC-E Accuracy

Dense

33.762846.501459.2471.9786May 15, 2026
Updated 16d ago

Evaluation Results

MethodLinks
2026.05
82.8360.2478.8272.8589.346.277.629043.1671.22
2026.05
81.1956.1478.9374.6678.2345.271.128940.168.29
2026.05
74.4946.2575.9968.977.7144.262.828739.6264.11
2026.05
67.6741.8163.6270.8870.2437.8770.048037.4259.95
2026.05
52.9933.7958.0166.3870.2835.853.437637.753.82
2026.05
49.8131.8349.2865.8273.733.857.047233.0151.81
2026.05
46.5133.8745.4567.2575.993377.266830.8153.13
2026.05
44.7827.3943.5753.9143.3930.651.996729.8643.61
2026.05
35.6528.9233.9160.4672.842948.386628.0444.8