Average (%): 45.5
Papers with Code · CC BY-SA 4.0
Multi-task Language Understanding
The test covers 57 tasks including elementary mathematics, US history, computer science, law, and more. https://arxiv.org/pdf/2009.03300.pdf
논문
Average (%): 39.2
Average (%): 87.0
Average (%): 47.9
Average (%): 53.2
Average (%): 52.6
Average (%): 44.8
Average (%): 33.6
Average (%): 45.3
MMLU (5-shot): 89.2
MiLe Loss: a New Loss for Mitigating the Bias of Learning Difficulties in Generative Language Models
Average (%): 29.68
Average (%): 55.0
Average (%): 37.9
Average (%): 28.0
Average (%): 49.9
Average (%): 39.2