56,966 blogs · 4,417,066 posts Claim your blog

Record · corpus.blog/posts/01a08765-eebf-7165-92d7-75802771988d

Dead Man's Switch: Combining SAE Features and Refusal Intervention in LLMs

cyberchris.xyz · published 18 September 2024

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at cyberchris.xyz Wayback capture

0blogs citing

First cited: —
Most recent: —
Rank this month: not ranked
Links from this post: 9
Words captured: 1,521

Cited by

No blog we hold has cited this post yet.

Links from this post

Link

Host

https://aisafetyfundamentals.com/

aisafetyfundamentals.com

https://huggingface.co/spaces/cyber-chris/dead-mans-switch

huggingface.co

https://github.com/cyber-chris/llm-dead-man-switch

github.com

https://neuronpedia.org/llama3-8b-it/25-res-jh/23610?embed=true&embedexplanation=true&embedplots=true&embedtest=true&height=300

neuronpedia.org

https://github.com/cyber-chris/llm-dead-man-switch/blob/main/scripts/tabulate_features.ipynb

github.com

Towards Monosemanticity: Decomposing Language Models With Dictionary Learning Read at the source ↗

transformer-circuits.pub

https://www.lesswrong.com/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector

lesswrong.com

https://www.neuronpedia.org/llama3-8b-it-res-jh

neuronpedia.org

https://jbloomaus.github.io/SAELens/

jbloomaus.github.io