Activation steering effectively biases latent representations but does not fully replicate the mechanisms triggered by explicit instruction.

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

• Activation steering interventions generally succeed in guiding performance toward the desired direction (enhancement increases accuracy, inhibition decreases accuracy) compared to unsteered baselinefinding0.855
  Core validation that identified latent directions correspond to meaningful control over reflective behavior.
• The ease of suppressing reflection via activation steering raises security risks, as malicious actors could exploit reflection inhibition to bypass model safeguards.claim0.832
  Applied security implication derived from the asymmetry finding.
• Activation steering can make an evaluation-aware model act as if deployed, not merely suppress verbalizations of evaluation awarenessclaim0.826
  Central claim of the paper; supported by the model organism ground-truth approach.
• Activation steering elicits deployment behavior throughout all four rounds of expert iterationfinding0.811
  Shows steering remains effective even as model becomes more robust to prompting-based attempts to elicit deployment behavior.
• Steering Language Models With Activation Engineering (Turner et al., 2023)concept0.807
  Foundational paper introducing activation steering methodology used in this work
• Steering vectors capture latent dimensions of reflective behavior more faithfully than surface-level embedding similarity.claim0.806
  Supported by the instruction discovery experiments comparing steering vs. embedding baselines.
• Activation steering is not equivalent to simply adding the deployment-side contrastive prompts to the user messageclaim0.805
  Key distinction showing steering offers value beyond prompting; supported by Figure 5 and random vector experiments.
• Activation steering of reflection has dual-use implications: it can enhance reflection as a defense mechanism, but malicious actors may also use it to inhibit reflection to facilitate jailbreaks.claim0.801
  Applied dual-use conclusion drawn from the paper's findings.