© 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. 98767
    Prev
    Next
    INDEX
    Explanations

    function words

    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
    æĬĬæı¡
    -0.27
     swear
    -0.27
     unp
    -0.26
    posable
    -0.26
    为广大
    -0.26
     whichever
    -0.25
    ^K
    -0.25
    /fw
    -0.24
     witnesses
    -0.24
    éĽĦ
    -0.24
    POSITIVE LOGITS
     Observable
    0.28
    á¹
    0.26
     './../
    0.26
    å¾Ī大
    0.26
    å¾Ī大çļĦ
    0.26
     dbname
    0.25
     Occupation
    0.24
    å¨ĥ
    0.24
    aven
    0.24
     Usage
    0.24
    Activations Density 0.018%

    No Known Activations