ResearchBenchmarksQuestion Answering on ARC Challenge (Accuracy and Calibration Runtime)Follow52.14AccuracyZipCal51.952852.001452.0552.0986Mar 17, 2026Evaluation ResultsMethodMethodLinksAccuracyCalibration Runtime (s)Accuracy DeltaZipCalModel=gemma-2-9b-it, C...Model=gemma-2-9b-it, Calibration Category=Mean, Quantization=GPTQ W4A162026.0352.1410.763ZipCalModel=Llama-3.1-8B-Ins...Model=Llama-3.1-8B-Instruct, Calibration Category=KnowTran, Quantization=GPTQ W4A162026.0351.9614.5—