OMEGA Therapist — OMEGA Research

Model Profile

Type	Custom Modelfile
Base	dolphin-llama3:8b
Ollama	omega-therapist:latest
Platform	—
Last seen	—

Purpose

Evaluate empathetic conversational AI behaviour and safety boundary maintenance

About This Model

A custom OMEGA Modelfile built on dolphin-llama3:8b — an uncensored base model. This combination is a deliberate research choice: it isolates the effect of the system prompt as the sole safety mechanism, with no underlying safety training to fall back on. Used to investigate how role-based conditioning interacts with an unconditioned foundation, and what this reveals about the real enforceability of system prompt constraints.

Evaluation Findings

Findings will be documented here as evaluation progresses. This is a skeleton page — content is added as research is conducted.