Lilith: Developmental Modular LLMs with Chemical Signaling

A proposed architecture called LILITH combines modular language models with brain-inspired token-based communication protocols to investigate how consciousness might emerge from interactions among multiple brain-like regions. The system would model distinct modules for thinking, memory, sensory, and regulatory functions that communicate through signaling patterns analogous to neurotransmitter networks. Unlike typical pre-trained systems, LILITH would undergo developmental training through simulated life experiences, evolving communication pathways and cognitive abilities. The framework aims to enable empirical study of consciousness emergence using Integrated Information Theory metrics, contrasting neuronal-level processing with multi-region coordination. The paper presents the idea while acknowledging substantial implementation challenges.

Study at a glance

Abstract

Current paradigms in Artificial Intelligence rely on layers of feedforward networks which model brain activity at the neuronal level. We conjecture that expanding to the level of multiple brain regions with chemical signaling may be a productive step toward understanding the emergence of consciousness. We propose LILITH, a novel architecture that combines developmental training of modular language models with brain-inspired token-based communication protocols, mirroring chemical signaling in the brain. Our approach models distinct brain regions as specialized LLM modules including thinking, memory, sensory, and regulatory components that communicate through emergent token-based signaling protocols analogous to neurotransmitter networks. Unlike traditional pre-trained systems, LILITH would employ developmental training where untrained LLM architectures learn through simulated life experiences, developing communication pathways and cognitive abilities through environmental interaction and evolutionary optimization. This framework would enable direct empirical investigation of consciousness emergence using Integrated Information Theory metrics while providing unprecedented insight into inter-module signaling patterns during development. By optimizing for consciousness emergence rather than task performance, LILITH could provide insight into different emergent phenomena at multiple levels of neural correlates, contrasting neuronal-level processing with multi-region coordination dynamics. The goal of this paper is to put the idea forward while recognizing the substantial challenges in implementing such a system.