From the course: Prompt Engineering with LangChain

Unlock the full course today

Join today to access over 24,600 courses taught by industry experts.

Agents: Let chains choose tools based on high-level directives

Agents: Let chains choose tools based on high-level directives

From the course: Prompt Engineering with LangChain

Start my 1-month free trial Buy for my team

Agents: Let chains choose tools based on high-level directives

“

- [Instructor] Agents in LangChain are systems that use a language model to interact with tools. they could be used for tasks such as grounded question answering, interacting with APIs, or taking action. LangChain provides a standard interface for agents, a selection of agents to choose from, and example of end-to-end agents. So one thing I want to drive home here is, I know it sounds fancy, but at the end of the day, agents are really nothing more than a prompting strategy. A sophisticated prompting strategy, but at the end of the day, agents are just prompts. So let's talk about the difference between agents and chains. So the core idea of an agent is to use an LLM to choose a sequence of actions. So in chains, a sequence of actions is hard coded through the prompt template, through code. But an agent, a language model, is going to be used as a reasoning engine to determine what actions to take and in which order. So an agent essentially is a language model that is interacting with…

Contents