Know a great AI tool or AI news source? Submit it to AiByts
LessWrong20 min read

Manifold Steering from a Natural State is Effective and Generalizable, but Relies on Model Accuracy

tl;dr I continue work to apply Goodfire lab's manifold steering methods to the Gemma-2-2b model's natural state, both in raw residual space and in the space produced by Anthropic's pre-trained CLT features.

By Anna MarbutLessWrong
Manifold Steering from a Natural State is Effective and Generalizable, but Relies on Model Accuracy

Read the full story

Originally published by LessWrong on Oct 12, 2026.

Read on LessWrong(opens in a new tab)

Comments

Sign in to join the discussion.

No comments yet — be the first to share your thoughts.