© 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. 20-TRANSCODER-HP
    7. 129882
    Prev
    Next
    INDEX
    Explanations

    Identifying problems and changes

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_20.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.20.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    .Commands
    -0.29
    arov
    -0.29
    ammen
    -0.27
    æĹłæĦı
    -0.26
    byn
    -0.26
    çİ°å®ŀä¸Ń
    -0.25
    ObjectId
    -0.24
    caa
    -0.23
    nek
    -0.23
    cko
    -0.23
    POSITIVE LOGITS
    èijĹ
    0.27
    aram
    0.27
    人åı£
    0.26
    ãĥ¼ãĤ·
    0.26
    ä»·éĴ±
    0.26
    åıij
    0.26
    ç¬Ļ
    0.25
    åĺ´åĶĩ
    0.25
    its
    0.24
    fter
    0.24
    Activations Density 47.784%

    No Known Activations