Anti-Persistence of Emotion Features

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

• PCs of the emotion space and persistenceconcept0.783
  Analysis showing that lower-rank (more central) PCs of emotion feature activations are more persistent than higher-rank (noisier) PCs
• Emotional State Persistenceconcept0.775
  The property of emotion features maintaining elevated activation well beyond the local token context that triggered them
• If persistence is genuinely related to emotion features, lower PCs of the emotion space (more central, less noisy) should be more persistent; if it is an artifact, noisier PCs should have similar persistence.hypothesis0.773
  Falsifiability test built into the PC analysis design
• Mechanism by which activation of an emotion feature sometimes leads to later suppression of that same featurequestion0.767
  Identified research gap: the paper observes anti-persistence but has no explanation for it
• To what extent is emotion feature persistence driven by genuine internal emotional state versus autoregressive conversational context dynamics?question0.764
  Core open question the paper raises but does not fully resolve
• autoregressive persistenceconcept0.761
  Baseline persistence of any probe direction arising from the autoregressive nature of LLMs, not specific to emotion content
• Self-evaluated emotionality and textual evaluation of SAE features predict persistence in opposite directions.claim0.759
  Surprising finding that the two evaluation methods diverge in their relationship with persistence
• steerable emotion featuresconcept0.759
  Emotion-encoding directions in LLM activation space that can be amplified or suppressed via activation steering to causally drive model behavior