Archetypal LLMs

Model adaptation

Language models shaped by organizational policy, decision logic, and structured archetypal feedback.

Archetypal LLMs investigates how language models can learn an organization's decision patterns and policy context while retaining useful capabilities. The documented direction combines open-weight model adaptation, organizational ontologies, persistent memory, and distinct reasoning roles. The research asks how to make specialization measurable: compare adapted and baseline models on held-out decisions, test memory retention and semantic drift, and evaluate collaborative reasoning without assuming that named personas improve outcomes.

Evidence priorities

Make progress measurable.

The initiative is framed around testable questions. Its research direction should be assessed against observed evidence and the limits of each experiment.

• Held-out organizational decision quality against an unchanged model baseline
• Retention of general capability and memory after adaptation
• Drift, auditability, and the contribution of each reasoning role

Research goals and proposed mechanisms are not established findings. Product-linked tracks describe the research behind a capability, rather than a claim that its hypotheses have been validated.