List your company

Zero-shot Question Answering on (PiQA, ARC-E, ARC-C, HellaSwag, Winogrande, BoolQ, OBQA, SiQA)

80.63PiQA AccuracyFull Rank
50.41858.261566.10573.9485Apr 1, 2026

Evaluation results

MethodLinks
Full RankModel=LLaMA3-8B, Compression Ratio=0%, Evaluation Protocol=Zero-shot2026.0480.6377.7453.579.1273.3281.1944.833.0165.41
Full RankModel=LLaMA2-7B, Compression Ratio=0%, Evaluation Protocol=Zero-shot2026.0479.0574.5446.2575.9968.977.7144.232.9162.44
OBD-LLMModel=LLaMA2-7B, Compression Ratio=20%, Evaluation Protocol=Zero-shot2026.0474.3758.6734.961.3466.2270.0336.234.1454.48
OBD-LLMModel=LLaMA3-8B, Compression Ratio=20%, Evaluation Protocol=Zero-shot2026.0473.7259.4734.959.4767.9668.5636.832.9654.23
SVD-LLMModel=LLaMA2-7B, Compression Ratio=20%, Evaluation Protocol=Zero-shot2026.0472.6352.6132.2558.6864.1764.3136.833.6251.88
SVD-LLMModel=LLaMA3-8B, Compression Ratio=20%, Evaluation Protocol=Zero-shot2026.0472.2554.8432.6857.5166.0668.293732.9152.69
ASVDModel=LLaMA2-7B, Compression Ratio=20%, Evaluation Protocol=Zero-shot2026.0463.8247.2229.6146.657.2259.630.431.9945.81
ASVDModel=LLaMA3-8B, Compression Ratio=20%, Evaluation Protocol=Zero-shot2026.0463.3843.8627.6541.9556.7566.4532.433.4245.73
FWSVDModel=LLaMA3-8B, Compression Ratio=20%, Evaluation Protocol=Zero-shot2026.0451.5825.2127.7326.2250.1247.429.233.5736.38
FWSVDModel=LLaMA2-7B, Compression Ratio=20%, Evaluation Protocol=Zero-shot2026.0451.5825.2127.7326.2250.1247.429.233.5736.38