Arditi and colleagues: a low-dimensional representation mediating model refusal

What changed

Identified a direction in residual-stream activations whose controlled intervention substantially changes refusal behavior in the tested instruction-tuned language models, linking an activation measurement to a causal behavioral effect.

Scientific scope

Identified a direction in residual-stream activations whose controlled intervention substantially changes refusal behavior in the tested instruction-tuned language models, linking an activation measurement to a causal behavioral effect. The result concerns selected prompts, intervention choices and thirteen tested models, not a theorem that every refusal in every model has one-dimensional cause. Refusal detection and harmful-response detection measure different outcomes.

Evidence

The linked original works support the scoped statement and contribution attribution. WRS has not independently replayed experiments or reconstructed a full mathematical proof. The result concerns selected prompts, intervention choices and thirteen tested models, not a theorem that every refusal in every model has one-dimensional cause. Refusal detection and harmful-response detection measure different outcomes.

Contributors

• The named authors in the research record · SCIENCE: Source-listed authors for the exact contribution; experimental subjects are not autonomous scientific authors. Source support and predecessor work retain separate roles. Individual effort shares and unresolved identities are not inferred.

• Refusal in Language Models Is Mediated by a Single Direction

Provisional WRS assessment, edition 2026-10-23.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction