List your company

Language Understanding on MMLU (Category Breakdown Scores)

37.31STEM Accuracy (MMLU)FP16
26.868429.579232.2935.0008Oct 16, 2024

Evaluation results

MethodLinks
FP16Backbone=LLaMA2-7B, Avg. Bit=16, Evaluation Protocol=five-shot2024.1037.3152.9143.0453.8246.46
CMPQBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=five-shot2024.1036.4850.5741.9152.1645
QuIPBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=five-shot2024.1036.2548.0739.8349.0743
AWQBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=five-shot2024.1036.2450.3141.7252.2244.85
RTNBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=five-shot2024.1034.5946.0539.6247.7841.83
GPTQBackbone=LLaMA2-7B, Avg. Bit=4, Evaluation Protocol=five-shot2024.1034.1343.8738.1345.5940.25
CMPQBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=five-shot2024.1032.5443.2636.3445.9939.27
AWQBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=five-shot2024.1031.7436.5930.8638.0933.98
GPTQBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=five-shot2024.1028.2425.6731.6932.9229.85
RTNBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=five-shot2024.1027.9323.6324.8724.1225.08
QuIPBackbone=LLaMA2-7B, Avg. Bit=3, Evaluation Protocol=five-shot2024.1027.2730.8127.0825.5127.57