A single direction encoding whether the model represents a claim as true or false is persona-relative across conversations (Lampinen et al. 2026)

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

• A linear preference vector encoding how much a model likes a given task is persona-relative: it activates strongly for creative writing in the assistant but for phishing in the evil personafinding0.830
  Evidence that core representations like preferences are persona-relative, supporting claim that personas gate content of representations
• Ududec et al. 2026 find that once a model has entered the evil persona region over the course of a conversation, it is difficult to steer it backfinding0.823
  Evidence that the evil persona region exhibits the stickiness hallmark of a genuine attractor basin
• The model-persona view entails that minds can have alien structure with simultaneous contradictory beliefs across parallel instances, but this is no stranger than humans having different beliefs at different timesclaim0.818
  Response to the main objection against the model-persona view about contradictory beliefs across simultaneous instances
• A model may produce internally coherent persona outputs while shifting characteristics across different generations, so intra-response and inter-response consistency must be assessed separatelyclaim0.803
  Supported by low correlation between ICatom and RCatom (r=0.44)
• Models are not merely tracking dialogue context features; same-concept steering shows privileged internal access is necessary to explain self-report shiftsclaim0.801
  Addresses skeptical alternative that reports reflect only conversational content
• Coding and writing conversations keep the model in the default Assistant persona range throughout, showing minimal driftclaim0.796
  Empirical characterization of conversation domains that are safe for model persona stability
• What if the concept being manipulated does not lie on a straight line in the model's representations?question0.795
  The motivating question that opens the paper and leads to the development of manifold steering.
• Hypothesis 1 (Gateway Features): Persona vectors act as gateway features—single directions in activation space that shape LLM behavior across most, if not all, contextshypothesis0.794
  First of three hypotheses about persona implementation in LLMs, motivating the persona views