© 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. 32-TRANSCODER-HP
    7. 42327
    Prev
    Next
    INDEX
    Explanations

    scientific and technical content

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_32.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.32.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    èĴĻ
    -0.32
    anner
    -0.28
    atten
    -0.26
    é±¼
    -0.26
     Hy
    -0.26
    atin
    -0.26
    GES
    -0.26
    æĻķ
    -0.25
    stadt
    -0.25
     PTR
    -0.25
    POSITIVE LOGITS
    osit
    0.31
    æĭĸ
    0.26
    kok
    0.26
    触åĬ¨
    0.24
    -drop
    0.24
     drifting
    0.24
    iped
    0.23
    éĺµéĺµ
    0.23
    è·ĸ
    0.23
    udiant
    0.23
    Activations Density 0.005%

    No Known Activations