Noising/Denoising Activation Patching

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Activation patchingmethod0.836
  Standard method in mechanistic interpretability that intervenes on activations; VPD flips this paradigm by patching parameters.
• Patch-Closureconcept0.757
  A set property meaning all coordinate patches of its elements remain within the set; proved equivalent to axis-aligned hyperrectangles
• Residual Stream Activation Patchingmethod0.750
  Used to localize causally implicated hidden states by swapping activations between true and false inputs
• Attribution patchingmethod0.743
  Gradient-based method to estimate the effect of zeroing a feature on a specific logit difference.
• Feature ablation (zeroing feature activations)method0.741
  Clamping a feature's value to zero to measure its causal effect on model output.
• Activationsconcept0.727
  Internal representations of the model on which probes operate; the method uses activations to rank datapoints.
• Activation Probingconcept0.725
  Technique of reading out model beliefs from internal activations before the final answer token is generated
• Activation decompositionconcept0.722
  The conventional approach (e.g., SAEs, transcoders) of decomposing activations into interpretable features.