alignmentforum.org/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector

Preview meta tags from the alignmentforum.org website.

Linked Hostnames

27

Thumbnail

Search Engine Appearance

Google

https://alignmentforum.org/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector

Steering GPT-2-XL by adding an activation vector — AI Alignment Forum

Alex Turner and collaborators show that you can modify GPT-2's behavior in surprising and interesting ways by just adding activation vectors to its f…



Bing

Steering GPT-2-XL by adding an activation vector — AI Alignment Forum

https://alignmentforum.org/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector

Alex Turner and collaborators show that you can modify GPT-2's behavior in surprising and interesting ways by just adding activation vectors to its f…



DuckDuckGo

https://alignmentforum.org/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector

Steering GPT-2-XL by adding an activation vector — AI Alignment Forum

Alex Turner and collaborators show that you can modify GPT-2's behavior in surprising and interesting ways by just adding activation vectors to its f…

  • General Meta Tags

    13
    • title
      Steering GPT-2-XL by adding an activation vector — AI Alignment Forum
    • title
      spotify-podcast-badge-wht-blk-165x40
    • charset
      utf-8
    • viewport
      width=device-width, initial-scale=1
    • Accept-CH
      DPR, Viewport-Width, Width

  • Open Graph Meta Tags

    5
    • og:title
      Steering GPT-2-XL by adding an activation vector — AI Alignment Forum
    • og:type
      article
    • og:url
      https://www.alignmentforum.org/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector
    • og:image
      https://res.cloudinary.com/lesswrong-2-0/image/upload/c_fill,ar_1.91,g_auto/SocialPreview/bj2r3jtoovu5gfcmpjrg
    • og:description
      Alex Turner and collaborators show that you can modify GPT-2's behavior in surprising and interesting ways by just adding activation vectors to its f…

  • Twitter Meta Tags

    3
    • twitter:image:src
      https://res.cloudinary.com/lesswrong-2-0/image/upload/c_fill,ar_1.91,g_auto/SocialPreview/bj2r3jtoovu5gfcmpjrg
    • twitter:description
      Alex Turner and collaborators show that you can modify GPT-2's behavior in surprising and interesting ways by just adding activation vectors to its f…
    • twitter:card
      summary

  • Link Tags

    32
    • alternate
      https://www.alignmentforum.org/feed.xml
    • canonical
      https://www.alignmentforum.org/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector
    • preload
      https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/Steering%20GPT-2-XL%20by%20adding%20an%20activation%20vector_a_compass_needle_adjusting_a_shi_0.26651304514537477/jqnevelzysz3adrrfefi
    • preload
      https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/FEGguDMzGbKojSQF9/pg9d78ympfbw0zygoqim
    • preload
      https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/FEGguDMzGbKojSQF9/eekt86qa7htfefs0sm5f

Links

153