A good parameter subcomponent is causally important only for specific roles and can be removed from the model without hurting performance on irrelevant prompts

{
  "doc": "paper.md",
  "context": "Definitional principle guiding VPD: subcomponents should encode narrow, targeted computational roles rather than distributed, multi-purpose functionality.",
  "category": "ai",
  "norm_label": "A good parameter subcomponent is causally important only for specific roles and can be removed from the model without hurting performance on irrelevant prompts",
  "graphify_id": "causal_importance_definition",
  "source_file": "paper.md",
  "imported_from": "/Users/antonborzov/Documents/Research.nosync/papers/extract_typed_out/paper/graph.json",
  "extracted_type": "claim",
  "source_location": "§1",
  "graphify_file_type": "claim"
}