A language model that forms persistent memories from conversation and maintains them through sleep. MEMIT weight editing + null-space-constrained maintenance.

An editable, auditable 807K-param byte-level LLM: CRUD single facts with provable per-edit locality, and abstain when unsure instead of guessing. CPU, offline.

This project investigates model editing techniques for Large Language Models (LLMs), focusing on how these models can be modified in specific domains without compromising overall performance.