method
active
method:prototype-contrast-loss-l-cePrototype Contrast Loss (L_CE)
Loss function pulling representations toward positive centroid and pushing away from negative centroid with angular margins
Neighborhood — ranked by edge-count
Papers (1)
paper
Frameworks (1)
framework
- ArcFace/CosFace Angular Margin LossimplementsAngular margin technique borrowed from face recognition and applied to the prototype contrast loss in SAE training
Methods (1)
method
- Procedure mapping hidden representations into SAE space and applying contrastive loss to learn facet-aligned control vectors
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.
- Auxiliary training objective from Grant (2025) that constrains intervened representations to remain near natural distribution
- Auxiliary objective combining L2 and cosine losses against pre-recorded CL vectors to improve causal relevance when one model is causally inaccessible.
- Regularization component of the composite loss that penalizes deviation from baseline model behavior on Alpaca instructions
- Distance-based loss comparing injected representation to class centroids in the active subspace
- One of two contrastive objectives analyzed; shown to be minimized by PMI kernel representation up to scaling
- Novel variant of CL loss introduced in this paper targeting only causal subspace dimensions to improve OOD performance
- The property that living structures contain intense contrast—far more than one imagines helpful; true opposites which annihilate each other when superimposed, creating differentiation that gives birth to something; contrast unifies rather than separates when used correctly
- The objective function combining L2 reconstruction error and L1 penalty scaled by decoder norm, used to train the SAE.