Latent Memory is a Module for Large Language Models that seek to integrate a vector-based memory system into the inference process, leveraging embeddings to capture deeper semantic meaning. - View it on GitHub