© 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. 89631
    Prev
    Next
    INDEX
    Explanations

    Steve

    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
    ôme
    -0.28
     appeal
    -0.28
    idad
    -0.28
    çīĩ
    -0.27
    cles
    -0.26
    çģ«çģ¾
    -0.25
    idade
    -0.24
    äºĭåĦ¿
    -0.24
    çīĩåŃIJ
    -0.24
     bor
    -0.24
    POSITIVE LOGITS
    å¾Ĺèµ·
    0.28
    opo
    0.28
     wrongly
    0.27
     sno
    0.24
    èµŀèµı
    0.24
     misunder
    0.24
     Xml
    0.23
    éĸ£
    0.23
     Williamson
    0.23
    .infinity
    0.23
    Activations Density 23.346%

    No Known Activations