© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 18-TRANSCODER-HP
    7. 88178
    Prev
    Next
    INDEX
    Explanations

    text fragments

    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
    Negative Logits
    èĩŃ
    -0.27
    èĢ½è¯¯
    -0.25
     delaying
    -0.25
    ÑĢаг
    -0.24
     delay
    -0.24
    çŁŃæļĤ
    -0.24
    纹
    -0.24
     Wholesale
    -0.24
    èħIJ
    -0.24
    enames
    -0.24
    POSITIVE LOGITS
    ...(
    0.26
     szer
    0.26
    ALAR
    0.25
    æĥ³æī¾
    0.25
    åĪĽç«ĭ
    0.25
    åıĮåıĮ
    0.25
    licable
    0.25
    æµĨ
    0.24
     nesting
    0.24
     analys
    0.24
    Activations Density 8.612%

    No Known Activations