Reinforcement learning (RL) algorithms have had tremendous success accounting for reward-based learning across species, including instrumental learning in contextual bandit tasks, and they capture ...