Decide what to do given some input.
Steps the LLM has taken so far, along with observations from each.
User inputs.
Optional
callbackManager: anyCallback manager.
Action specifying what tool to use.
Return response when agent has been stopped due to max iterations
Optional
_callbackManager: any
Class representing a single action agent using a LLMChain in LangChain. Extends the BaseSingleActionAgent class and provides methods for planning agent actions based on LLMChain outputs.
Example