NOTE: An early implementation of a project to implement the techniques described in this post can be found here. For those involved in vulnerability management, you will understanding how challenging it can be to triage the multitude of vulnerabilities that are raised by a variety of scanners. Some of them, despite their low severity, can be critical to a particular codebase due to how the affected functions are being used....
Custom LLM Agent With ChatGPT and LLaMA 2
Langchain is one more of the popular agents or frameworks for Large Language Models (LLMs). However, it can be extremely complicated to use, and as it is relatively mature, it can abstract away significant efforts that go into developing an LLM agent. For that purpose, Infuser was born. It was conceived for my learning; I wanted to build a ReAct agent without using Langchain. I started with ChatGPT as it was easier to implement; all I really needed was an API key....