When does the model initiate reflection during its reasoning process?

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

• How does reflection influence the model's reasoning performance?question0.835
  Second central research question motivating ReflCtrl investigation
• what is the underlying mechanism of self-reflection in reasoning LLMs?question0.797
  Open question motivating the entire paper; identified as not yet well understood
• Reflection is not merely a behavioral artifact of prompting but a phenomenon encoded in the model's activation space.claim0.792
  Central interpretive claim of the paper, supported by steering vector experiments.
• Reflective reasoning requires late-stage integration of semantic and reasoning signals, hence reflection-related directions emerge more clearly in higher network layers.claim0.783
  Interpretive claim about the locus of reflection in transformer architecture.
• Does the model internally maintain a form of 'consistency score' or probability mass over coherent reasoning trajectories, and how is this score modulated during reflection?question0.779
  Promising future research direction about the internal mechanism of error detection.
• The underlying mechanism of self-reflection in reasoning LLMs is not yet well understoodquestion0.776
  Broad gap motivating the entire paper
• Reflection does not only emerge in SFT or RL stages but arises earlier during pre-training.claim0.771
  Cited finding from Shah et al. contextualizing the training origins of reflection.
• a model becomes strongly confident in its final answer, but continues generating tokens without revealing its internal beliefquote0.766
  Core definitional quote for performative chain-of-thought