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

Natural Language Understanding on ARC-c

65.8Accuracy

mPLUG-Owl2

Updated 22d ago

Evaluation Results

Method	Links
mPLUG-Owl2 2023.11		65.8
Vicuna-v1.5 2023.11		56.6
LLaMA-2-Chat 2023.11		54.9
WizardLM 2023.11		47.5
SpaceByte + SP 2026.05		41.9
Fixed (p = 4) + SP 2026.05		41.6
H-Net + SP 2026.05		40.9
Fixed (p = 16) + SP 2026.05		40.8
Tokenizer-based 2026.05		40.7
Entropy-based 2026.05		40.6
Entropy-based + SP 2026.05		40.4
LLaMA-2 2023.11		40.3
H-Net 2026.05		40.1
Fixed (p = 4) 2026.05		39.8
Fixed (p = 8) + SP 2026.05		39.8
Fixed (p = 8) 2026.05		39.6
Byte-level 2026.05		38.7
SpaceByte 2026.05		38.6
Fixed (p = 16) 2026.05		35.2
ProDiaL 2024.11		30.8
FT 2024.11		29.61
ProDiaL 2024.11		29.61
FT 2024.11		29.52
LoRA 2024.11		29.52
FT 2024.11		29.27
Full-FT 2024.11		28.84
ProDiaL 2024.11		28.84
DoRA 2024.11		28.75
Strong 2024.11		28.67
DoRA 2024.11		28.67
LoRA 2024.11		28.54
DoRA 2024.11		28.5
LoRA 2024.11		28.41
BitFit 2024.11		26.88