AI brief
A LessWrong post discusses steering AI agent behavior, citing Anthropic's Golden Gate Claude experiment, in which a feature linked to the Golden Gate Bridge was extracted with a sparse autoencoder and clamped.
Written by AI from LessWrong's published text.
Background
From our own funding, model and news trackers, not from the source.
Anthropic
- Latest funding
- Growth · $65B · 28 May 2026
- Newest model on the API
- Claude Haiku 5.5 · 7 Oct 2026 · $0.1 in / $0.5 out per million tokens
Source
“Background The Linear Representation Hypothesis [1] has allowed for many remarkable steering experiments.”
Interpretation or community commentary rather than straight reporting.
Read original story ↗