List your company

Zero-shot Evaluation on Evaluation Suite (ARC, BoolQ, HellaSwag, PIQA)

69.82Average ScoreBF16
67.594468.172268.7569.3278May 29, 2025

Evaluation results

MethodLinks
BF16Bits=16, Finetuning=none, Incoherence processing=true, Backbone=Llama 3.1 8B Instruct2025.0569.8251.3778.0382.0557.7479.92
YAQA-ABits=4, Finetuning=none, Incoherence processing=true, Quantizer=INT4, Backbone=Llama 3.1 8B Instruct2025.0569.1150.6878.1179.6557.1379.98
YAQA-BBits=4, Finetuning=none, Incoherence processing=true, Quantizer=INT4, Backbone=Llama 3.1 8B Instruct2025.0568.9249.4977.3181.0156.9879.82
LDLQBits=4, Finetuning=none, Incoherence processing=true, Quantizer=INT4, Backbone=Llama 3.1 8B Instruct2025.0567.995076.9477.0156.8379.16
DISCQUANTBits=4, Finetuning=none, Incoherence processing=true, Quantizer=INT4, Backbone=Llama 3.1 8B Instruct2025.0567.6850.3477.4474.1256.5579.92