artifact
active
artifact:probe-based-data-attribution-landing-page-goodfire-ai-research-probe-based-data-attribution

Probe-Based Data Attribution Landing Page (goodfire.ai/research/probe-based-data-attribution)

Web landing page for the research paper hosted by Goodfire.

Neighborhood — ranked by edge-count

Methods (1)

method
  • Probe-Based Data Attribution
    about
    Linear classifier approach applied to model activations to identify which training datapoints caused undesired behaviors in post-training.