finding
active
finding:base-specific-features-include-llm-roleplay-and-smartphone-assistant-dialogue-featuresBase-specific features include LLM roleplay and smartphone assistant dialogue features
Interactions at odds with Claude's trained behavior.
Source paper
extracted_fromRelated 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 strong claim that there is no underlying authentic agent behind the simulator, only layers of role play
- Conditional prediction about how a well-informed dialogue agent would handle questions of personal identity
- Claude 3 Opus ratings aligned with human judgment of feature descriptions.
- E.g. 'Here is a draft', apology, and 'explain like I'm in sixth grade' features.
- Primary empirical claim of the paper
- Empirical characterization of conversation domains that are safe for model persona stability
- Novelty claim establishing the paper's contribution relative to prior work focused on closed-form tasks