LLM Powered Autonomous Agents
Highlights
- ▸Weng maps LLM agents to three components: planning (subgoal decomposition + self-reflection), memory (short-term in-context learning + long-term vector retrieval), and tool use (external API calls) — the LLM is the brain, the infrastructure is the body
- ▸Tree of Thoughts (Yao et al. 2023) extends Chain of Thought into an active search tree with BFS or DFS traversal, evaluating each reasoning state by classifier or majority vote — not linear reasoning but branching search
- ▸LLM+P (Liu et al. 2023) delegates long-horizon planning to classical PDDL planners instead of the LLM itself — the agent translates natural language to PDDL, the planner executes, and the LLM translates results back
Original excerpt
Building agents with LLM (large language model) as its core controller is a cool concept. Several proof-of-concepts demos, such as AutoGPT, GPT-Engineer and BabyAGI, serve as inspiring examples. The potentiality of LLM extends beyond generating well-written copies, stories, essays and programs; it can be framed as a powerful general problem solver.
In a LLM-powered autonomous agent system, LLM functions as the agent’s brain, complemented by several key components:
Planning * Subgoal and decomposition: The agent breaks down large tasks into smaller, manageable subgoals, enabling efficient handling of complex tasks. * Reflection and refinement: The agent can do self-criticism and…
10 more articles in this vault.
Import the full Lilian Weng vault to Burn 451 and build your own knowledge base.
Content attributed to the original author (Paul Graham). Burn 451 curates publicly available writing as a reading index. For removal requests, contact @hawking520.