search

iKala / ievals

Official github repo for TMMLU+, Large scale traditional chinese massive multitask language understanding
MIT License
43 stars 2 forks source link

Groq, Reka, Together API support #5

Closed theblackcat102 closed 3 months ago

theblackcat102 commented 5 months ago
model_name STEM other (business, health, misc.) social sciences humanities Average
meta-llama/Llama-3-70b-chat-hf 34.44 39.51 47.02 37.50 39.62
meta-llama/Llama-3-8b-chat-hf 31.52 31.79 34.19 28.91 31.60
google/gemma-7b-it 31.89 33.79 35.70 34.00 33.84