Agentic Discovery & Exploration with gpt-4o-mini

I have been reading quite a bit or research on AI Agents, and specially have experimented notebooks from LangChain and LlamaIndex (Agentic RAG) using gpt-4o-mini. I find the WebVoyager research paper particularly interesting together with the LangChain implementation. And using the gpt-4o-mini model is more than sufficient…

Any other resources someone can point me to on the topic of AI Agents?

First off agents used with AI predate the modern notion of LLM agent.

In my view of agents, which predates most of what is now being noted, is to just think of an agent as a subtask that uses AI to accomplish a specific task. In other words if a monolithic prompt does not work, try breaking the process into subtask using a different prompt(s). That is basically it.

When I read your question it is like asking in more traditional programming for information on how to create a method or function. It is such an expansive question that one really needs to ask something more specific.

HTH


If you insist on a place to look then might I suggest

Search | arXiv e-print repository


Bonus info:

Sometimes people publish survey papers that cover a topic at a higher level, basically a quick read to become familiar with a topic or to bring one’s knowledge up to date.

Here is the same query but listing just the survey papers (survey in the title of the paper) which is a reasonably sized list.

Search | arXiv e-print repository

3 Likes

Thanks for the insights Eric.

1 Like