© 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. 35-TRANSCODER-HP
    7. 39673
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_35.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.35.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ople
    -0.27
    ergic
    -0.25
     Blo
    -0.25
    utures
    -0.25
    /cli
    -0.24
    績
    -0.24
    blo
    -0.24
    ergy
    -0.23
    åĭĭ
    -0.23
    åı¸
    -0.23
    POSITIVE LOGITS
    æĪijè¿ĺæĺ¯
    0.26
    大éĻĨ
    0.25
     OVERRIDE
    0.25
    lin
    0.24
    elight
    0.24
    onen
    0.23
    æĪIJé¾Ļ
    0.23
    reads
    0.23
    ält
    0.22
    noinspection
    0.22
    Activations Density 19.157%

    No Known Activations

    This feature has no known activations.