Our new X account is live! Follow @wizwand_team for updates

Common Sense Reasoning on HellaSwag (test)

83.9Accuracy

Llama-3.3-70B-Base

Updated 4d ago

Evaluation Results

Method	Links
Llama-3.3-70B-Base 2026.02		83.9
MISTRAL-7B 2026.01		83.43
QWEN3-14B 2026.01		81.44
LLAMA-3.1-8B 2026.01		79.31
Mistral-3.2-24B-Base 2026.02		79.3
FP16 2025.12		78.5
Gemma-3-27B-Base 2026.02		78.2
Gemma-3-12B-Base 2026.02		77.7
Apertus-70B-Base 2026.02		77.4
OLMo-3-32B-Base 2026.02		77.2
Qwen-3-30B-A3B-Base 2026.02		76.5
Qwen-3-14B-Base 2026.02		76.2
DEEPSEEK-R1-QWEN-8B 2026.01		75.72
Llama-3.1-8B-Base 2026.02		75.7
DEEPSEEK-R1-LLAMA-8B 2026.01		74.78
CALM 2025.12		74.5
LLAMA-3.2-3B 2026.01		74.13
DART 2026.01		74.12
QWEN3-4B 2026.01		73.75
EuroLLM-9B-Base 2026.02		73.6
EuroLLM-22B-Base 2026.02		73.2
Apertus-8B-Base 2026.02		73.2
SpinQuant 2025.12		72.1
SmoothQuant 2025.12		71.8
FP16 2025.12		70.2
AWQ 2025.12		69.5
OLMo-3-7B-Base 2026.02		69.2
DART 2026.01		68.4
CALM 2025.12		66.5
GPTQ 2025.12		65.2
DART 2026.01		64.58
SpinQuant 2025.12		64.1
SmoothQuant 2025.12		63.5
AWQ 2025.12		60.2
DART 2026.01		59.49
DART 2026.01		58.63
GPTQ 2025.12		55.8
DART 2026.01		53.66
DART 2026.01		52.77
Dip-SVD 2026.02		49
SAES-SVD 2026.02		47.7
SVD-LLM 2026.02		47
SAES-SVD 2026.02		47
Dip-SVD 2026.02		40.2
SVD-LLM 2026.02		35.5