Unverified50% confidenceFactExact time
WRDO (Weighted Refusal Direction Optimization) is a method that combines each prompt-specific refusal direction into a composite direction through learned aggregation weights.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe same optimization mechanism of WRDO can be used not only in the forward direction to weaken refusal, but also in reverse to restore refusal capability in model checkpoints from which refusal behavior has been removed.73% similarUnverifiedBilinear interpolation is differentiable with respect to offsets, allowing gradients of sampling positions to be propagated through backpropagation, which is the mathematical foundation for end-to-end training of deformable attention.62% similarUnverifiedWaypoint-OPD gates the endpoint shift based on direction consistency: when the direction is consistent it is retained (gate value close to 1), and when the direction conflicts it is suppressed (gate value close to 0).59% similarUnverifiedThe Abliteration technique identifies and suppresses the direction vector responsible for refusal behavior in the model's residual stream, using orthogonal projection to remove that direction from the weight matrices, thereby eliminating refusal behavior without retraining.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1084577API
curl https://kongchang.com/api/v1/knowledge/claims/1084577MCP
get_claim(id=1084577)