model_name: "RedHatAI/Llama-3.2-1B-Instruct-quantized.w8a8"
accuracy_threshold: 0.31
num_questions: 1319
num_fewshot: 5
server_args: "--enforce-eager --max-model-len 4096"