© 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. 25-TRANSCODER-HP
    7. 107669
    Prev
    Next
    INDEX
    Explanations

    Informal text

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_25.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.25.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    tribute
    -0.29
    riers
    -0.28
    oleon
    -0.27
    posal
    -0.27
     por
    -0.26
     oral
    -0.25
    apolis
    -0.25
    æºIJäºİ
    -0.25
    furt
    -0.25
    ::$_
    -0.25
    POSITIVE LOGITS
    Drupal
    0.27
    flo
    0.25
    _PULL
    0.24
    åĪĹ表
    0.24
     flo
    0.24
    WithContext
    0.24
    Constants
    0.24
    æĻĥ
    0.24
     metabol
    0.24
    åĴĮæĬĢæľ¯
    0.24
    Activations Density 0.930%

    No Known Activations