© 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. 30-TRANSCODER-HP
    7. 96581
    Prev
    Next
    INDEX
    Explanations

    application scaling

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Cree
    -0.31
    \Services
    -0.27
     equip
    -0.26
    éĮ
    -0.25
    ä¿¡æģ¯æľįåĬ¡
    -0.25
     Ps
    -0.24
    æĽ´æľī
    -0.24
     Sug
    -0.24
     Clar
    -0.24
    idences
    -0.24
    POSITIVE LOGITS
     scale
    0.35
     SCALE
    0.35
    _scale
    0.34
     Scale
    0.33
    Scale
    0.32
     scales
    0.31
    scale
    0.30
    Scaler
    0.29
    elize
    0.29
     management
    0.29
    Activations Density 0.895%

    No Known Activations