© 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. 88309
    Prev
    Next
    INDEX
    Explanations

    Named entities and 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
    matcher
    -0.30
    tings
    -0.28
     Nimbus
    -0.27
    ammen
    -0.26
    èĥŀ
    -0.26
    _dur
    -0.26
    &r
    -0.26
    liner
    -0.25
    uja
    -0.25
    /settings
    -0.25
    POSITIVE LOGITS
    æłĩ
    0.27
    ĵį
    0.26
    çŃīå½¢å¼ı
    0.25
    -CN
    0.25
    chen
    0.24
    ngen
    0.24
    łģ
    0.24
    éĸ£
    0.24
     Do
    0.23
    mailto
    0.23
    Activations Density 12.628%

    No Known Activations