LLM Personalization

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

• LLM Self-Correctionconcept0.817
  Related capability where LLMs correct their own outputs, studied via linear representations.
• LLM Internal Representationsconcept0.798
  High-dimensional vectors produced at each transformer layer for each input token; the primary substrate analyzed in this study.
• LLM Introspective Self-Reportconcept0.787
  The capacity of Kimi K2.5 to evaluate its own internal emotional state when steered, used as a novel interpretability signal
• Persona Assignment to LLMsconcept0.787
  The practice of providing LLMs with a persona description to shape their generated responses
• Linear Representation of Concepts in LLMsconcept0.780
  The finding that interpretable concepts including character traits are encoded as linear directions in transformer residual streams
• LLM psychosisconcept0.776
  Tendency for models to get lost in roleplay or doom spirals, mitigated by expanded awareness.
• Reflection in LLMsconcept0.773
  The core phenomenon studied: the ability of LLMs to evaluate and revise their own reasoning.
• LLM-Judge Data Attributionmethod0.772
  Alternative data attribution approach using an LLM as a judge; compared against the probe-based method.