© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 20-TRANSCODER-HP
    7. 63198
    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
    Not in Any Lists

    No Comments

    Negative Logits
     Orth
    -0.30
    Orth
    -0.27
    acer
    -0.27
    /owl
    -0.27
     Swe
    -0.26
    åħļåĴĮ
    -0.26
    {/
    -0.26
     orth
    -0.26
    -touch
    -0.25
    ç«Ļ
    -0.24
    POSITIVE LOGITS
    ä¸ĭåİ»
    0.34
    ä¸ĭ游
    0.32
     downstream
    0.31
    ä¸ĭæĿ¥
    0.30
    ìĿij
    0.27
    -down
    0.27
    rips
    0.26
     ve
    0.24
    .setInput
    0.24
    iola
    0.24
    Activations Density 54.972%

    No Known Activations

    This feature has no known activations.