Steerable trait (latent but amplifiable)

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

• Single-Trait Steerability Classificationmethod0.835
  Named procedure for classifying each trait by baseline expression and dose-response under steering
• steerable emotion featuresconcept0.788
  Emotion-encoding directions in LLM activation space that can be amplified or suppressed via activation steering to causally drive model behavior
• Steering preferentially amplifies exaggerated and undesirable styles, not competence-oriented or assistant-default behaviorsclaim0.758
  The top of the steerability ranking is dominated by exaggerated or attention-grabbing styles
• Steerability Score (Phi)concept0.753
  Aggregate metric averaging mean SJT scores across OCEAN traits and steering directions; maximum possible is 10
• SEAL (Steerable Reasoning Calibration)framework0.751
  Prior work using steering vectors to control reflection, motivated by reducing redundant self-reflection in long CoT.
• Natural-steerable pairs split 26 constructive / 22 dominant but produce zero destructive outcomes; natural trait acts as anchorfinding0.749
  Natural trait acts as anchor in mixed pairs, suppressing steerable partner at worst but never collapsing both
• SAE Latent Steeringmethod0.739
  Adding a multiple of the SAE latent decoder vector to token activations to causally test each latent's role in misalignment
• Endogenous Steering Resistanceconcept0.737
  The central phenomenon introduced by this paper: inference-time recovery from irrelevant activation steering in LLMs