© 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. 26037
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "and" as well as phrases that imply action or intent

    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
    .moveTo
    -0.19
     moved
    -0.18
     move
    -0.17
     moves
    -0.17
    moved
    -0.17
     Moved
    -0.17
    åīįå¾Ģ
    -0.17
     travelled
    -0.17
    移åĬ¨
    -0.17
    move
    -0.16
    POSITIVE LOGITS
     there
    0.19
     begin
    0.18
     där
    0.18
     beginning
    0.18
    éĤ£éĩĮ
    0.18
     dort
    0.17
     ÑĤам
    0.17
    begin
    0.17
     Begins
    0.17
     daar
    0.17
    Activations Density 0.043%

    No Known Activations