Evidence that altercentric biases in a continuous false belief task depend on highlighting the agent's belief