Colorful visual signals can provide receivers with valuable information about food, danger, and the quality of social partners. However, the value of the information that color provides varies depending on the situation, and color may even act as a sensory trap where signals that evolved under one context are exploited in another. Despite some elegant early work on color as a sensory trap, few empirical studies have examined how color biases may vary depending on context and under which situations biases can be overridden. Here, using Neolamprologus pulcher, a highly social cichlid fish from Lake Tanganyika, we conducted a series of experiments to determine color biases and investigate the effects of these biases under different contexts. We found that N. pulcher interacted the most with yellow items and the least with blue items. These biases were maintained during a foraging-based associative learning assay, with fish trained using yellow stimuli performing better than those trained using blue stimuli. However, these differences in learning performance did not extend to reversal learning; fish were equally capable of forming new associations regardless of the color they were initially trained on. Finally, in a social choice assay, N. pulcher did not display a stronger preference for conspecifics whose yellow facial markings had been artificially enhanced. Together, these findings suggest that the influence of color biases varies under different contexts and supports the situational dependency of color functions.