Positive Experience Reflection for Agents in Interactive Text Environments
This addresses a specific bottleneck in interactive text environments for AI agents, but it is incremental as it builds on existing reflection methods.
The paper tackles the problem of agents in text-based games struggling with reduced effectiveness after initial success or when using smaller LLMs, and introduces Sweet&Sour, which improves performance by incorporating positive experiences and managed memory.
Intelligent agents designed for interactive environments face significant challenges in text-based games, a domain that demands complex reasoning and adaptability. While agents based on large language models (LLMs) using self-reflection have shown promise, they struggle when initially successful and exhibit reduced effectiveness when using smaller LLMs. We introduce Sweet&Sour, a novel approach that addresses these limitations in existing reflection methods by incorporating positive experiences and managed memory to enrich the context available to the agent at decision time. Our comprehensive analysis spans both closed- and open-source LLMs and demonstrates the effectiveness of Sweet&Sour in improving agent performance, particularly in scenarios where previous approaches fall short.