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

Commonsense Reasoning on WinoGrande (0-shot, test)

73.95Accuracy

FP16

Updated 1mo ago

Evaluation Results

Method	Links
FP16 2025.12		73.95
AQLM-1x16 + PV-Tuning 2025.12		70.24
AQLM-2x8 + PV-Tuning 2025.12		69.14
CodeGEMM-m1v4g128 + PV-Tuning 2025.12		69.06
AQLM-1x16 2025.12		68.75
CodeGEMM-m2v8g128 + PV-Tuning 2025.12		68.75
Qwen3-8B + CAT-Q 2026.06		65.43
TernaryLLM-8B+KD 2026.06		65
TequilaLLM-3B 2026.06		62.7
Qwen3-4B + CAT-Q 2026.06		62.19
TriLM-3.9B 2026.06		62.1
Qwen3-4B + CAT-Q 2026.06		61.96
CodeGEMM-m2v8g128 2025.12		60.54
BitNetV1-3.9B 2026.06		60.5
BitNetV1-7B 2026.06		59.98
CodeGEMM-m1v4g128 2025.12		58.96
AQLM-2x8 2025.12		58.25
TriLM-1.5B 2026.06		56.1
BitNetV1-1.3B 2026.06		55.8
FlexRound-q2g128 2025.12		55.16
Qwen3-1.7B + CAT-Q 2026.06		54.83