© 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. 19273
    Prev
    Next
    INDEX
    Explanations

    Code/paths/filenames

    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
    Pager
    -0.30
    vais
    -0.27
    .ARR
    -0.27
    æĴ¬
    -0.26
    ascade
    -0.26
    ahan
    -0.26
     paddle
    -0.26
    _handle
    -0.26
    uarios
    -0.25
    (handle
    -0.24
    POSITIVE LOGITS
    ane
    0.29
    ĮĢ
    0.27
    å¥īçĮ®
    0.27
     Counseling
    0.26
    åĬŁ
    0.25
    ç¥ĿæĦ¿
    0.25
    å¤©åľ°
    0.24
     partitions
    0.24
    uz
    0.23
    decl
    0.23
    Activations Density 0.003%

    No Known Activations