© 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. 19219
    Prev
    Next
    INDEX
    Explanations

    Snippets from articles

    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
    riters
    -0.27
    åĬĽåѦ
    -0.27
    _frag
    -0.27
    elements
    -0.26
    centre
    -0.26
    ãĤ¹ãĥĿ
    -0.25
    é¢ģ
    -0.25
    licity
    -0.24
    örper
    -0.24
    åľ¨ä¼łå¥ĩ
    -0.24
    POSITIVE LOGITS
    ä¿ĿæĮģ
    0.29
    æĹ¥å¸¸å·¥ä½ľ
    0.27
     subscription
    0.27
    èĹı
    0.26
     office
    0.26
     daily
    0.26
    ews
    0.26
    æĪij们çļĦ
    0.25
     our
    0.25
    ç¶Ń
    0.25
    Activations Density 0.886%

    No Known Activations