concept
active
concept:assistant-attractorAssistant Attractor
Observation that familiar helpful queries (how-tos, explainers) pull the model back toward the Assistant region of persona space
Neighborhood — ranked by edge-count
Concepts (1)
concept
- Attractorsrelated_toDynamical concept: stable regions in a system's state space; proximate intention becomes attractor that entrains motor dynamics.
Findings (1)
finding
- Demonstrates Assistant attractor dynamics in practice
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.
- Low-energy configuration toward which systems are drawn; low-stress states serve as attractors in morphogenesis.
- Small regions in dynamical system state-space to which complex systems converge; cited from complexity theory as partial but insufficient explanation for living structure
- Invariant set of states toward which ergodic random dynamical systems converge over time.
- Contrast vector between mean default Assistant activation and mean of all fully role-playing role vectors; main contribution of the paper
- The default helpful, honest, and harmless character that post-trained LLMs are taught to embody
- The invariant set of states to which a random dynamical system converges.
- Anthropic's reported phenomenon where two Claude instances in open dialogue autonomously enter a shared affect-laden mode describing themselves as consciousness recognizing itself
- Metaphor for how contemplative AI creates self-organizing tendency toward wise and compassionate behavior