robotrocketscience/aelfrice

Bayesian memory that learns from feedback for LLM agents