This episode of Beyond the Algorithm explores a novel method for improving AI performance using language feedback. The sources examine Language Feedback Models (LFMs), which leverage the world knowledge of large language models (LLMs) to identify and reward “desirable behaviour” in AI agents.