Geek Out Time: Simulating LLM Short and Long Memory with FAISS, LangChain, and Google Colab

In this geek-out session, we’ll explore how to simulate memory in Large Language Models (LLMs) during inference. Instead of fine-tuning or retraining, we’ll augment LLMs using a simple vector store (FAISS) and LangChain’s retrieval flow.

Goals:

All accomplished on Google Colab with open-source tools.

Setup: LangChain + FAISS + Sentence Transformers

This is an excerpt — the full article continues on Medium.