© 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. Qwen3-4B
    5. 20-TRANSCODER-HP
    7. 108736
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_20.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.20.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    kim
    -0.27
     fit
    -0.27
    kor
    -0.27
     strokeLine
    -0.27
    igit
    -0.26
    tpl
    -0.26
    ç§Ģ
    -0.26
    SIGN
    -0.25
    åĨĴ
    -0.25
    coln
    -0.25
    POSITIVE LOGITS
    _RB
    0.27
    <header
    0.26
    躲
    0.25
    (push
    0.25
    lder
    0.24
    èĢ³
    0.24
    UserCode
    0.23
    æĬĬä½ł
    0.23
    ceeded
    0.23
    åĿļæĮģ
    0.23
    Activations Density 3.830%

    No Known Activations

    This feature has no known activations.