© 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. 18-TRANSCODER-HP
    7. 90215
    Prev
    Next
    INDEX
    Explanations

    and

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_18.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.18.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    azzo
    -0.27
    echa
    -0.27
     DI
    -0.26
     hack
    -0.26
    DI
    -0.26
     stmt
    -0.25
     scrutiny
    -0.25
    antz
    -0.25
    stmt
    -0.24
     Cont
    -0.24
    POSITIVE LOGITS
    åĭĴ
    0.30
    lev
    0.29
    ella
    0.28
    ful
    0.27
    解éĶģ
    0.26
    nbsp
    0.26
    åĴĮçĿ¦
    0.25
    kus
    0.25
    æľīæĽ´å¥½çļĦ
    0.25
    è³´
    0.25
    Activations Density 16.739%

    No Known Activations