Line of inquiry
Inquiring lines›How do architecture and constraint…›What determines model reasoning ca…›this line of inquiry
Can activation steering reliably reveal and control model reasoning?
A broader line of inquiry — a family of 41 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 41
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- Can causal steering change verbalization without changing internal representation?
- Does this reasoning steering method work consistently across all model sizes?
- Can activation-space steering vectors replicate thinking model performance without retraining?
- Can activation steering causally control evaluation framing effects across downstream tasks?
- Can targeted activation steering surface latent reasoning in base models?
- Can activation patching reveal which reasoning steps actually matter?
- Can steering vectors prove that representations are genuinely organized?
- Can activation steering directly steer models toward concise reasoning without prompting?
- Can activation-space interventions reach biases that prompting cannot address?
- Can interventions on model components prove mechanism without explaining encoding?
- Can we steer model reasoning by manipulating single features?
- Can activation steering vectors compress reasoning without retraining models?
- Can activation steering compress reasoning without retraining models?
- How do probe-based interventions in activation space compare to mechanistic interpretability approaches?
- Do reading vectors from activation space causally control model behavior?
- Does causal intervention alone explain how neural mechanisms implement representations?
- How do prompting and activation steering relate as compression strategies?
- Can steering internal features bypass or override prompt-level instructions in simulations?
- What causes gradient-based steering via natural language descriptions to work?
- Does base model geometry predict which associations persist through intervention?
- Can steering vectors be combined with other compression techniques?
- What makes some concepts more steerable than others in activation space?
- Can steering reshape the capabilities and safety split without changing eval-awareness rates?
- Can activation-space directions reliably steer LLM behavior without retraining or prompting?
- Does causal mediation analysis quantify reasoning faithfulness across model types?
- Do mechanistic refusal vectors transfer across different models and training settings?
- Can agents revise their beliefs predictably when presented with interventions?
- How does motivational stage determine which interventions actually work for users?
- What training architecture models the causal structure of partner influence?
- Can negative feedback through critiques achieve the same steering flexibility as positive preferences?
- Why do steered pain directions disable consequence-weighting but preserve factual competence?
- Can you steer reasoning by directly manipulating SAE features?
- Can causal models be extended to include non-causal cognition?
- How do causal belief networks extracted from interviews enable intervention reasoning?
- Can architectural changes reorder when uncertainty and empowerment signals influence decisions?
- Can causal belief networks extracted from interviews predict how people respond to policy changes?
- How can extracted causal belief networks enable intervention simulation?
- How do we measure the cognitive flow cost of different intervention strategies?
- Can belief networks from interviews simulate how people change their minds?
- How does the prefrontal cortex inspire artificial reasoning architectures?
- Why should low-probability severe risks trigger early intervention?