© 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. 22-TRANSCODER-HP
    7. 87207
    Prev
    Next
    INDEX
    Explanations

    guarding

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_22.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.22.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Suspension
    -0.30
    sticks
    -0.26
    Å¡ÃŃ
    -0.25
    nesc
    -0.25
    ä¸İåIJ¦
    -0.24
     sensed
    -0.24
    sed
    -0.24
    å°ijåIJĥ
    -0.24
    -Shirt
    -0.24
    æ·±åħ¥å¼Ģå±ķ
    -0.23
    POSITIVE LOGITS
    ames
    0.27
     triple
    0.26
    inger
    0.25
    ulation
    0.24
    urous
    0.24
    ingly
    0.24
    -Identifier
    0.24
     means
    0.24
    icularly
    0.23
    adamente
    0.23
    Activations Density 0.030%

    No Known Activations