LLM Wiki
Overview
A centralized knowledge base for tracking developments, architectures, and tools related to Large Language Models, with a focus on local deployment, memory systems, and agent capabilities.
Recent Developments
Persistent Memory Systems
Current research and implementations are shifting from static context windows to dynamic, persistent memory structures for local AI agents.
- Librarian System Architecture: A new approach to providing local AI with evolving, persistent memory.
- Source: Developing Persistent, Intelligent Memory for Local AI with a Librarian System
- Key Concept: Moves beyond simple context injection by implementing a structured “librarian” system that manages memory retrieval and storage, allowing the AI to maintain state across sessions without relying solely on repeated prompt feeding.
- Reference: Developing Persistent, Intelligent Memory for Local AI with a Librarian System (Codacus, 2026-07-13)