© 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. 26217
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate biased or extreme perspectives in discussions

    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
    ieten
    -0.17
    agi
    -0.17
    rek
    -0.15
    ãģĭãĤı
    -0.15
     ONLY
    -0.15
    _only
    -0.15
     Only
    -0.14
    .configure
    -0.14
    reater
    -0.14
    raÄį
    -0.14
    POSITIVE LOGITS
     most
    0.41
     slightest
    0.33
     smallest
    0.33
     simplest
    0.28
    -most
    0.27
    most
    0.26
    æľĢ
    0.24
     strongest
    0.23
     Most
    0.21
    Most
    0.21
    Activations Density 0.090%

    No Known Activations