Using the wake/sleep manipulation, we investigate the impact of sleep on evaluative learning and counter-bias training. In the evaluative learning experiment, participants completed a judgement task and a memory task for two times in a 12-hour interval. Their judgement responses were fitted to the RCB model by Heycke and Gawronki (2020). Their memory recall were scored into five types to separate the memory difference of individual information and integrated information. In three studies focusing on counter-bias training, participants completed gender-STEM implicit association test at pre-training, post-training and 12-hour delay sessions. Their D scores were calculated to compare the effect of group and time. We also fit the data to Quad model by Conrey et al. (2005) to examine the underlying mental processes of bias reduction.