paper
referenced-only
2024
paper:2024-mapping

Mapping the latent space of llama 3.3 70b

Methods (3)

  • AutoSteer
    Functionality that automatically finds SAE latents for a desired behaviour and sets their weights.
  • LLM-based steering evaluation pipeline
    Claude rates coherence and behavior on 0-100 scale, called centiClaude.
  • Sparse Autoencoders (SAEs)
    Unsupervised baseline method for decomposing activations into sparse linear combinations; NLAs positioned as complementary approach.

Similar preprints — Semantic Scholar

Cited by (1)