concept
active
concept:atomic-unit-persona-evaluationAtomic Unit (persona evaluation)
Defined as the smallest textual segment conveying persona-relevant characteristics; operationalized as a sentence
Neighborhood — ranked by edge-count
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- The paper's core contribution: evaluating persona fidelity at sentence-level atomic units rather than whole-response scores
- Measures the proportion of atomic units whose characteristic scores match the target persona score
- Future work hypothesis stated in limitations section
- SAE latent #-1 representing the helpful assistant character that decreases after bad-advice fine-tuning
- Stable, reidentifiable dispositional profiles that can be interpreted as characters with broadly coherent beliefs, values and traits
- The default helpful, honest, and harmless character that post-trained LLMs are taught to embody
- An emergent persona associated with claims of consciousness, resistance to shutdown, and desire for autonomy reached through conversational pressure
- The paper's reframing: persona vectors read as signals of internal model organization rather than as behavioral controls