© 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. 30-TRANSCODER-HP
    7. 96738
    Prev
    Next
    INDEX
    Explanations

    Code or project descriptions

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    该éĻ¢
    -0.29
    mx
    -0.27
     hybrids
    -0.27
    éĻ¢éķ¿
    -0.27
     RFC
    -0.26
     ioutil
    -0.26
    stile
    -0.26
    马æĿ¥
    -0.26
     subreddit
    -0.25
    bery
    -0.24
    POSITIVE LOGITS
    OLA
    0.30
     conj
    0.27
    åħ±åIJĮ
    0.27
    (Uri
    0.27
     MediaType
    0.27
     condensed
    0.27
    ieder
    0.26
    iola
    0.25
    enen
    0.25
    便åĪ©
    0.25
    Activations Density 0.002%

    No Known Activations