Emotion Features in LLMs

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

• Emotion geometry in LLM activationsconcept0.830
  Emotion emerges early, peaks in middle layers, sharpens with scale, and persists across tokens in LLM activations per Zhang & Zhong 2025
• Are LLM emotion states encoded only selectively in token positions where they are operative, or in a more complex non-linear manner?question0.802
  Question raised by Anthropic and partially addressed by this paper's persistence evidence
• Emotion features in LLMs are genuinely more persistent than variance-matched random features, indicating stateful emotional encoding beyond autoregressive dynamicsclaim0.800
  Central interpretive claim of the paper supported by multiple convergent analyses
• We hypothesize that persistently active emotional state representations exist in LLMs but may be missed by standard probing methods.hypothesis0.793
  Open hypothesis from the Anthropic paper that motivates this work
• Emotive states in LLMsconcept0.792
  Directions in activation space associated with contrastive emotive concept pairs studied in this paper as targets for introspection
• Linear Representation of Concepts in LLMsconcept0.783
  The finding that interpretable concepts including character traits are encoded as linear directions in transformer residual streams
• To what extent is there persistence of emotional state beyond what is expected merely from the autoregressive nature of LLMs?question0.775
  The central research question motivating the paper
• emotion feature persistenceconcept0.757
  The phenomenon that emotion feature activations remain elevated above baseline beyond local token bursts, measurable as long-range correlation