© 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. 18-TRANSCODER-HP
    7. 89650
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    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
    Not in Any Lists

    No Comments

    Negative Logits
     convo
    -0.29
    åı¯æĢľ
    -0.26
    ään
    -0.26
     Benefit
    -0.25
    -cookie
    -0.25
    sti
    -0.25
    毳
    -0.24
    ç©´
    -0.24
    ï¼Įé»ĺ认
    -0.24
    é¢Ĺ
    -0.23
    POSITIVE LOGITS
    缮
    0.27
    askell
    0.25
    ilde
    0.25
     travels
    0.25
    çľ¼ä¸Ń
    0.24
    orde
    0.24
    èĮ¼
    0.24
    èĸª
    0.23
    mdir
    0.23
    basename
    0.23
    Activations Density 5.828%

    No Known Activations

    This feature has no known activations.