Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore Recursive Language Models (RLMs), an advanced inference technique that enables large language models to interact with arbitrarily long prompts through external REPLs in this comprehensive 50-minute tutorial. Learn how RLMs allow language models to write code for exploring, decomposing, and transforming prompts while recursively invoking sub-agents to complete smaller subtasks, with subagent responses returned as symbols or variables rather than being automatically loaded into the parent agent's context. Examine actual RLM trajectories on real problems to understand their practical applications and see step-by-step implementation from scratch using Deno and Pyodide. Discover the key features and benefits of RLMs, including when and why to use this powerful technique for building sophisticated AI agents. Access the accompanying GitHub repository and PyPI package for hands-on experimentation with the fast-rlm implementation, and gain insights into this cutting-edge approach that represents a significant advancement in language model inference capabilities.
