© 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. 22-TRANSCODER-HP
    7. 97451
    Prev
    Next
    INDEX
    Explanations

    Licenses, copyright, and menu

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_22.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.22.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    .po
    -0.29
    PEAR
    -0.28
    po
    -0.27
    æŃ£ç¡®
    -0.26
    ä»·
    -0.26
    åĨĻ
    -0.25
     SU
    -0.25
    è¿ĩäºĨ
    -0.25
    hb
    -0.24
    SU
    -0.24
    POSITIVE LOGITS
     mans
    0.29
     dick
    0.26
    inding
    0.25
    ä¸Ĭåįĩ
    0.25
    adow
    0.24
    ardi
    0.24
    çĤ·
    0.24
     Mansion
    0.24
     chaired
    0.24
    /desktop
    0.23
    Activations Density 0.007%

    No Known Activations