Does instructing the model to assess correctness affect the geometry of truth directions?

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

• We hypothesize that explicitly instructing the model to evaluate the correctness of the given statement may change the geometry of truth directions.hypothesis0.911
  Motivating hypothesis for Section 5's investigation of prompt template effects.
• What is the effect of model instructions on truth directions?question0.847
  Research question motivating Section 5.
• Accuracy does not vary linearly with latent reflection directions; instead it follows a more non-linear mapping that requires deeper theoretical treatment.claim0.804
  Theoretical limitation identified by the authors distinguishing reflection from stylistic tasks.
• Are the discovered truth directions robust to architectural variation and fine-tuning differences across model families?question0.804
  Open question on generalization beyond Gemma and Qwen families
• Multiple semantically adjacent truth directions make models more vulnerable to manipulations that shift outputs without obvious signs in the primary truth directionclaim0.792
  Safety implication derived from multi-dimensional truth structure finding
• Linear truth directions in LLMs are reliable primarily in factual recall cases and break down when truth assessment depends on computing and storing intermediate results.claim0.790
  Central empirical conclusion of the paper about the fundamental limits of truth directions.
• Where inside the LLM should we look for an accurate truth direction that will generalize the most across tasks?question0.786
  One of the three guiding research questions of the paper.
• Discovered truth directions are highly specific and do not interfere with general instruction-following behaviorclaim0.785
  Interpretation of KL divergence retention results