finding
active
finding:policy-reduction-is-10-from-the-policy-itself-being-less-hallucinatoryPolicy Reduction is ~10% from the policy itself being less hallucinatory
Component of decomposition.
Source paper
extracted_from(2026) · Aaditya Vikram Prasad · Connor Watts · Jack Merullo · Dhruvil Gala +3
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Topline result.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Rules out overcautiousness.
- Decomposition component.
- The reduction in hallucination rate under inline interventions occurs due to a change in model beliefs in-context.hypothesis0.747Mechanism hypothesis for future work.
- Train-time scaling.
- 56% of test-time hallucinations are caught; of these 22% Fixed, 36% Retracted, 42% improperly handledfinding0.739Fig. 8 decomposition.
- Implication of PRH: larger models should amplify bias less and hallucinate less if they better model reality
- Demonstrates practical utility of preventative steering in a realistic deployment scenario
- What mechanisms underlie the changes to the policy, including the reduction when interventions are inlined?question0.722Open interpretability question.