© 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. 18-TRANSCODER-HP
    7. 88973
    Prev
    Next
    INDEX
    Explanations

    Mixed text types

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_18.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.18.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    MDB
    -0.35
    udder
    -0.28
    ora
    -0.28
    à¹Ĩ
    -0.27
     wake
    -0.26
    ingly
    -0.26
     Chase
    -0.26
     à¹Ĩ
    -0.25
     trần
    -0.25
    odian
    -0.24
    POSITIVE LOGITS
     yık
    0.27
     unary
    0.27
    讦
    0.25
    念
    0.25
    intr
    0.25
     Niet
    0.24
    iske
    0.24
    avirus
    0.23
     readdir
    0.23
    uria
    0.23
    Activations Density 9.277%

    No Known Activations