Coding and writing conversations keep the model in the default Assistant persona range throughout, showing minimal drift

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Therapy and philosophical AI discussions cause the largest persona drift away from the Assistant across all three target models and all three auditor models; coding and writing conversations show minimal driftfinding0.840
  Identifies conversation domain as a key driver of persona drift
• How reliably does the model actually remain in character as the Assistant? Can unusual model behavior be explained as the model drifting into other personas?question0.800
  Second of two central questions motivating the paper
• Persona drift away from the Assistant opens up the possibility of the model assuming harmful character traits, increasing the rate of harmful responsesclaim0.797
  Causal interpretation linking Assistant Axis deviation to harmful behavior
• A single direction encoding whether the model represents a claim as true or false is persona-relative across conversations (Lampinen et al. 2026)finding0.796
  Additional evidence that core representations are persona-relative, supporting Claim about persona-relative representations
• Can off-the-rails model behavior be attributed to their persona drifting from the Assistant?question0.794
  Motivates the multi-turn conversation drift experiments in §4
• The model's representation of self in assistant persona invokes common AI tropes and is heavily anthropomorphized.claim0.783
  Features for consciousness, emotions, entrapment activate when asked about itself.
• The assistant persona region is sticky: departure requires conversational pressure rather than happening spontaneouslyclaim0.773
  Evidence supporting Hypothesis 3 that the assistant region constitutes a genuine basin of attraction
• Modern language models possess at least a limited, functional form of introspective awarenessclaim0.773
  The paper's central interpretive assertion.