© 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. 89485
    Prev
    Next
    INDEX
    Explanations

    Musical notes

    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
     <<↵
    -0.28
    “My
    -0.26
     Perr
    -0.25
    æĺİçĻ½
    -0.24
    “In
    -0.24
    âĢ¦)↵↵
    -0.24
     *@
    -0.23
    çļĦ人æĿ¥è¯´
    -0.23
    clo
    -0.23
    âĢ¦”↵↵
    -0.23
    POSITIVE LOGITS
    æĮĩæķ°
    0.31
    åĪĿ次
    0.27
    ISO
    0.27
     subscript
    0.26
    boxed
    0.26
     rated
    0.25
    -grade
    0.25
     fed
    0.25
    adi
    0.25
     worth
    0.24
    Activations Density 9.746%

    No Known Activations