Towards Developmentally Plausible Rewards: Communicative Success as a Learning Signal for Interactive Language Models

Open in new window