Commonsense Reasoning on MMLU

78.95Accuracy

IAP-mv

Updated 10d ago

Evaluation Results

Method	Links
IAP-mv 2024.09		78.95
IAP-ss 2024.09		77.07
Prompt #5 2024.09		76.84
Prompt #4 2024.09		74.21
AMV (#1-7) 2024.09		63.86
Prompt #2 2024.09		61.93
IAP-mv 2024.09		59.65
Prompt #7 2024.09		57.19
Prompt #6 2024.09		56.67
IAP-ss 2024.09		56.49
Prompt #1 2024.09		55.79
Prompt #6 2024.09		55.25
IAP-mv 2024.09		54.39
Self-discover 2024.09		52.63
Prompt #5 2024.09		52.63
AMV (all) 2024.09		52.53
AMV (all) 2024.09		52.46
Prompt #4 2024.09		52.11
Prompt #2 2024.09		50.35
Prompt #2 2024.09		50.18
Prompt #4 2024.09		48.95
Prompt #3 2024.09		48.25
Prompt #5 2024.09		46.67
Prompt #3 2024.09		44.56
Prompt #3 2024.09		43.51
Prompt #9 2024.09		43.33
Prompt #1 2024.09		42.48
Prompt #7 2024.09		42.28
AMV (#1-7) 2024.09		41.23
In-Place TTT 2026.04		37.42
Prompt #1 2024.09		37.02
In-Place TTT 2026.04		36.48
Full Attention 2026.04		36.43
Sliding-Window Attention (SWA) 2026.04		36.06
Prompt #6 2024.09		35.79
Prompt #9 2024.09		30.7
Prompt #8 2024.09		26.32
Prompt #8 2024.09		22.56
Prompt #9 2024.09		20.18
Prompt #7 2024.09		14.56
Prompt #8 2024.09		9.3