© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 21104
    Prev
    Next
    INDEX
    Explanations

    references to drinks and beverages, particularly concerning alcohol

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Negative Logits
    åŃĺäºİ
    -0.18
     Bott
    -0.17
     Bottle
    -0.17
     bottle
    -0.16
    lint
    -0.16
     bott
    -0.15
    utor
    -0.15
    assis
    -0.15
    erial
    -0.15
     bottles
    -0.15
    POSITIVE LOGITS
     drinks
    0.38
     drink
    0.38
     beverages
    0.33
     beverage
    0.32
     Bever
    0.32
     Drink
    0.32
     Drinks
    0.31
    Drink
    0.29
    drink
    0.29
     liquid
    0.28
    Activations Density 0.013%

    No Known Activations