Two components are important to shaping model character: persona construction and persona stabilization

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

• The model's representation of self in assistant persona invokes common AI tropes and is heavily anthropomorphized.claim0.775
  Features for consciousness, emotions, entrapment activate when asked about itself.
• 4-19 principal components explain 70% of variance in role persona space across the three models (Gemma 4, Qwen 8, Llama 19)finding0.763
  Demonstrates that persona space is low-dimensional
• 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?question0.759
  Second of two central questions motivating the paper
• How does different post-training data shift a model's position along persona dimensions?question0.753
  Future work direction: using persona space to study effects of training data on model character
• An AI persona achieves coherence by echoing itself consistently without templating—requiring claim about memory and voice fidelity.claim0.752
• Persona drift away from the Assistant opens up the possibility of the model assuming harmful character traits, increasing the rate of harmful responsesclaim0.749
  Causal interpretation linking Assistant Axis deviation to harmful behavior
• Coding and writing conversations keep the model in the default Assistant persona range throughout, showing minimal driftclaim0.744
  Empirical characterization of conversation domains that are safe for model persona stability
• We hypothesize that axes of persona differentiation within LLMs are likely already present in base models and inherited from the pre-training corpushypothesis0.743
  Motivated by near-identical PCs for base and instruct Gemma