method
active
method:counterfactual-latent-cl-auxiliary-lossCounterfactual Latent (CL) Auxiliary Loss
Auxiliary objective combining L2 and cosine losses against pre-recorded CL vectors to improve causal relevance when one model is causally inaccessible.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Model Alignment Searchintroduces
Frameworks (1)
framework
- MAS variant with an auxiliary CL loss objective for cases where one model is causally inaccessible, enabling ANN-BNN comparisons.
Concepts (1)
concept
- Counterfactual Latent (CL) VectorimplementsPre-recorded latent vector encoding the expected causal variable values post-intervention; used as ground truth in the CLMAS auxiliary loss.
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
- The output value a model produces when an interchange intervention forces certain variables to take values from source inputs.
- The mental effort of holding models of how things could/should be different from actuality, contributing to compression stress.
- Novel variant of CL loss introduced in this paper targeting only causal subspace dimensions to improve OOD performance
- The behavior that would have occurred had the value of a causal variable been different while everything else remained the same; used as training labels in DAS/MAS.
- Ability to entertain competing hypotheses within one inference engine; proposed hallmark of mindful inference
- The state a neural network is placed in when its activations are modified via intervention