List your company

Zero-shot Performance on LMEval Harness Subset (HellaSwag, BoolQ, WinoGrande, PiQA, ARC)

75.46Mean AccuracyGIFT-SW
40.079249.264658.4567.6354Aug 27, 2024Nov 22, 2024Feb 18, 2025May 17, 2025Aug 13, 2025Nov 9, 2025Feb 5, 2026

Evaluation results

MethodLinks
GIFT-SWBits=4 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0875.46——————
STEBits=4 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0875.29——————
STEBits=4 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0874.84——————
GIFT-SWBits=4 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0874.5——————
STEBits=3 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0874.37——————
GIFT-SWBits=3 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0874.34——————
FP16Model=L3-8B, Bits=16, Group size=1282026.0274.3173.3280.979.2381.3177.7453.33
QUIK + LORABits=4 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0874.27——————
GIFT-SWBits=3 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0873.27——————
SNRQ-SModel=L3-8B, Bits=4, Group size=128, Q.Time (s)=498.72026.0273.2673.2480.2377.8880.1876.7651.26
GPTAQModel=L3-8B, Bits=4, Group size=128, Q.Time (s)=703.22026.0273.2373.6980.3478.1679.9976.4850.7
FP16Model=L2-13B, Bits=16, Group size=1282026.0273.2372.380.5279.3880.5877.5349.06
SNRQ-CModel=L3-8B, Bits=4, Group size=128, Q.Time (s)=495.22026.0273.0873.0980.2777.8379.9476.6250.72
LDLQModel=L3-8B, Bits=4, Group size=128, Q.Time (s)=425.12026.0273.0573.1680.1278.1479.5876.6750.61
GPTQModel=L3-8B, Bits=4, Group size=128, Q.Time (s)=4892026.0272.9873.5180.1977.9780.287649.95
SNRQ-CModel=L2-13B, Bits=4, Group size=128, Q.Time (s)=726.12026.0272.872.3880.1178.6579.8676.9548.82
LDLQModel=L2-13B, Bits=4, Group size=128, Q.Time (s)=623.92026.0272.7972.5879.8778.7579.9876.7948.75
GPTAQModel=L2-13B, Bits=4, Group size=128, Q.Time (s)=1019.22026.0272.7171.8779.8378.880.0976.8848.82
SNRQ-SModel=L2-13B, Bits=4, Group size=128, Q.Time (s)=709.42026.0272.6772.0679.9978.7579.7976.6948.75
GPTQModel=L2-13B, Bits=4, Group size=128, Q.Time (s)=727.72026.0272.6272.0679.9178.7879.4276.8348.72
GIFT-SWBits=4 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0872.53——————
STEBits=4 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0872.43——————
QUIK + LORABits=3 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0871.65——————
QUIK + LORABits=3 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0871.3——————
SNRQ-SModel=L2-13B, Bits=3, Group size=128, Q.Time (s)=724.22026.0271.1171.1979.0775.9578.8574.9446.63
QUIK + LORABits=4 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0871.08——————
SNRQ-CModel=L2-13B, Bits=3, Group size=128, Q.Time (s)=749.62026.0271.0670.6979.4476.0378.3375.1846.72
GIFT-SWBits=3 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0871——————
GPTAQModel=L2-13B, Bits=3, Group size=128, Q.Time (s)=10122026.027170.8379.3676.3679.0574.1146.42
LDLQModel=L2-13B, Bits=3, Group size=128, Q.Time (s)=719.22026.0270.9770.2679.1176.2478.9474.846.47
GPTQModel=L2-13B, Bits=3, Group size=128, Q.Time (s)=820.52026.0270.9370.5479.2776.3279.0873.9146.45
FP16Model=L2-7B, Bits=16, Group size=1282026.0270.4769.0679.1175.9877.7474.6246.33
STEBits=3 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0870.24——————
STEBits=3 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0869.82——————
SNRQ-SModel=L2-7B, Bits=4, Group size=128, Q.Time (s)=405.62026.0269.6569.278.5875.1476.4173.4745.07
GPTQModel=L2-7B, Bits=4, Group size=128, Q.Time (s)=481.62026.0269.5869.1478.475.1177.2573.1344.47
LDLQModel=L2-7B, Bits=4, Group size=128, Q.Time (s)=404.82026.0269.5469.6978.4475.2176.7773.0444.11
SNRQ-CModel=L2-7B, Bits=4, Group size=128, Q.Time (s)=450.92026.0269.4968.8378.5675.0776.273.4944.81
GPTAQModel=L2-7B, Bits=4, Group size=128, Q.Time (s)=610.52026.0269.4369.7278.5275.1675.3273.344.57
SNRQ-SModel=L3-8B, Bits=3, Group size=128, Q.Time (s)=484.72026.0269.3370.6477.5673.2877.5571.6245.36
GIFT-SWBits=2 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0867.61——————
SNRQ-CModel=L3-8B, Bits=3, Group size=128, Q.Time (s)=492.32026.0267.670.4876.4373.0676.3667.5441.76
SNRQ-SModel=L2-7B, Bits=3, Group size=128, Q.Time (s)=417.32026.0267.3468.0277.471.6975.0770.1841.37
GPTQModel=L3-8B, Bits=3, Group size=128, Q.Time (s)=547.52026.0266.8269.1475.5273.0773.7667.142.32
SNRQ-CModel=L2-7B, Bits=3, Group size=128, Q.Time (s)=407.42026.0266.8267.4877.6871.5572.9470.0241.26
LDLQModel=L2-7B, Bits=3, Group size=128, Q.Time (s)=427.82026.0266.566.8477.0271.7374.1268.4940.79
GPTAQModel=L2-7B, Bits=3, Group size=128, Q.Time (s)=590.42026.0266.3466.7177.277273.168.1140.84
GPTQModel=L2-7B, Bits=3, Group size=128, Q.Time (s)=467.92026.0266.1465.9776.7171.8873.1167.9341.21
GPTAQModel=L3-8B, Bits=3, Group size=128, Q.Time (s)=7152026.0265.7769.6674.773.8974.7561.8939.76
LDLQModel=L3-8B, Bits=3, Group size=128, Q.Time (s)=472.52026.0265.269.8274.9770.8171.2763.4740.84
QUIK + LORABits=4 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0863.99——————
QUIK + LORABits=3 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0862.91——————
STEBits=2 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0862.19——————
GIFT-SWBits=2 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0861.09——————
GIFT-SWBits=2 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0858.89——————
STEBits=2 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0858.2——————
QUIK + LORABits=2 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0853.8——————
STEBits=2 bit, Backbone=LLaMA3-8b, Fine-tuning dataset=TÜLU-V2-mix2024.0848.96——————
QUIK + LORABits=2 bit, Backbone=LLaMA2-13b, Fine-tuning dataset=TÜLU-V2-mix2024.0847.14——————
QUIK + LORABits=2 bit, Backbone=LLaMA2-7b, Fine-tuning dataset=TÜLU-V2-mix2024.0841.44——————