pulsatrix 1.1.0
Loading...
Searching...
No Matches
pulsatrix::LatentScaling Struct Reference

Latent Scaling (Minder et al., Eq. 4): how much of latent j's b-side decoder direction each source actually needs. For a target Y^m over the inputs where it's computed, β^m = ⟨Y^m d, f⟩ / (|f|² |d|²) with d = d_j^b, the best scale of f d against Y^m. More...

#include <crosscoder.hpp>

Public Attributes

int64_t latent = 0
 
double beta_error_a = 0
 
double beta_error_b = 0
 
double beta_reconstruction_a = 0
 
double beta_reconstruction_b = 0
 
double nu_error = 0
 
double nu_reconstruction = 0
 
int64_t active = 0
 Inputs where the latent fired.
 

Detailed Description

Latent Scaling (Minder et al., Eq. 4): how much of latent j's b-side decoder direction each source actually needs. For a target Y^m over the inputs where it's computed, β^m = ⟨Y^m d, f⟩ / (|f|² |d|²) with d = d_j^b, the best scale of f d against Y^m.

  • Error targets are what the rest of the crosscoder leaves: x^m - (x̂^m - f_j d_j^m). A b-only latent whose ν_error is high is complete shrinkage: source a needs it too, but the sparsity penalty zeroed its a decoder.
  • Reconstruction targets are x̂^m. A high ν_reconstruction is latent decoupling: the same thing is also written to source a by other latents.
  • ν = β^a / β^b. Minder et al. call a latent b-specific when ν_reconstruction < 0.5 and ν_error < 0.2; a negative β^b leaves ν meaningless, and they drop such latents.

Member Data Documentation

◆ active

int64_t pulsatrix::LatentScaling::active = 0

Inputs where the latent fired.

◆ beta_error_a

double pulsatrix::LatentScaling::beta_error_a = 0

◆ beta_error_b

double pulsatrix::LatentScaling::beta_error_b = 0

◆ beta_reconstruction_a

double pulsatrix::LatentScaling::beta_reconstruction_a = 0

◆ beta_reconstruction_b

double pulsatrix::LatentScaling::beta_reconstruction_b = 0

◆ latent

int64_t pulsatrix::LatentScaling::latent = 0

◆ nu_error

double pulsatrix::LatentScaling::nu_error = 0

◆ nu_reconstruction

double pulsatrix::LatentScaling::nu_reconstruction = 0

The documentation for this struct was generated from the following file: