What limitations prevent decoding strong truth directions?

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

• The need for genuine counting over lists of more than two elements introduces the key limitation of truth directions.claim0.772
  Identified as the exact computational operation that breaks truth direction generalization.
• Discovered truth directions are highly specific and do not interfere with general instruction-following behaviorclaim0.764
  Interpretation of KL divergence retention results
• Universality claims for truth directions are more limited than previously assumed, with significant differences observable for various model layers, task difficulties, task types, and prompt templates.claim0.757
  Overarching conclusion summarizing the paper's contribution relative to prior universality claims.
• Are the discovered truth directions robust to architectural variation and fine-tuning differences across model families?question0.754
  Open question on generalization beyond Gemma and Qwen families
• Truth directions fail to generalize to harder tasks (F3-F5) regardless of prompt template because activations remain highly entangled.claim0.749
  Establishes task difficulty as a hard limit that instructions cannot overcome.
• What is the effect of model instructions on truth directions?question0.747
  Research question motivating Section 5.
• Where inside the LLM should we look for an accurate truth direction that will generalize the most across tasks?question0.737
  One of the three guiding research questions of the paper.
• Truth directions emerge in earlier layers for factual tasks and later layers for arithmetic tasks.claim0.736
  Core empirical claim about the layer-dependence of truth direction emergence as a function of task type.