The Hierarchical Intrinsically Motivated Agent (HIMA) is an algorithm that is intended to exhibit an adaptive goal-directed behavior using neurophysiological models of the neocortex, basal ganglia, and thalamus.

Biologically inspired language model using Jaccard Surprise as its only training signal. No backprop. No GPU. Online Hebbian learning from corrections. Two-layer cortex with apical feedback. Runs on CPU under 200MB.

Long-term memory for LLM agents using Sparse Distributed Representations — no vector DB, no GPU, no external server.