© 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. 31-TRANSCODER-HP
    7. 18609
    Prev
    Next
    INDEX
    Explanations

    code imports/packages

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     lookahead
    -0.30
     mph
    -0.28
    æł¡
    -0.27
    æĵĤ
    -0.26
    个å¤ļ
    -0.26
    å¢ħ
    -0.25
     Len
    -0.25
    .cor
    -0.24
     Thr
    -0.24
    =top
    -0.24
    POSITIVE LOGITS
    ingle
    0.30
     visual
    0.28
     scrapy
    0.27
     Caj
    0.27
     visualize
    0.26
     Ced
    0.26
    agens
    0.26
    ãģ¾ãĤĮ
    0.26
     Visual
    0.26
     Cord
    0.24
    Activations Density 0.002%

    No Known Activations