LLM Chronicles #6.4: LLM Agents with ReAct (Reason + Act)

In this episode we’ll cover LLM agents, focusing on the core research that helped to improve LLMs’ reasoning while allowing them to interact with the external world via the use of tools. These include Chain of Thought prompting, PAL (Program-aided Language Models) and ReAct (Reason + Act) as used in Langchain and CrewAI agents.
Series website: llm-chronicles...
🖹 Canvas:
llm-chronicles...
🕤 Timestamps:
00:13 - Table of Contents
01:23 - Chain of Thought Prompting
03:10 - PAL (Program-aided Language Models)
05:14 - ReAct (Reason + Act)
09:22 - Tools, Plugins, Functions, APIs
10:54 - ReAct in Practice (JSON/XML formats, fine-tuned models)
12:05 - Function Calling (OpenAI)
13:08 - Modified ReAct (Browser agents, CodeAct)
14:15 - Summary
14:47 - Limitations & Cyber Security Considerations
References:
Chain-of-Thought Prompting Elicits Reasoning in Large Language Models, arxiv.org/abs/...
Large Language Models are Zero-Shot Reasoners, arxiv.org/abs/...
PAL: Program-aided Language Models, arxiv.org/abs/...
ReAct: Synergizing Reasoning and Acting in Language Models, arxiv.org/abs/...
InternLM: github.com/Int...
Executable Code Actions Elicit Better LLM Agents, arxiv.org/pdf/...
OpenDevin CodeACT: xwang.dev/blog...

[@laithvideos1]

Really appreciative of how you have condensed a complicated topic into 15 minutes. I must have gone to 20 other videos from other experts till I found this golden nugget. You are the best. Would be great to see practical use cases from for Agentic workflow like the customer one you mentioned.

[@donatocapitella]

Thank you for the support, it makes a big difference for small channels like this one! In September I'm planning 1 or 2 lab videos on agentic workflows, the second will likely be the implementation of a full agent, such as the customer assistant example.

[@MarkKelly76]

Great explanation. Thanks for producing this content.

[@donatocapitella]

Glad it was helpful!

[@henrylee_hd]

Excellent explanation about ReAct

[@Galaxy-wm7qr]

Great explanation

[@donatocapitella]

@@Galaxy-wm7qr thank you!!

[@cmthimmaiah]

Superb, so well explained. Appreciate it

[@donatocapitella]

@@cmthimmaiah thanks so much for the support 🙏

[@micbab-vg2mu]

Great video - thank you:)

[@KhapitarBalakAagya]

I watched all the videos of your channel, it has great content, and the amount of energy you put to make viewer to understand things is appreciable at utmost respect.

[@donatocapitella]

Thank you! The trick is that I'm trying to make myself understand the stuff as well! :)