© 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. 20-TRANSCODER-HP
    7. 141099
    Prev
    Next
    INDEX
    Explanations

    Mixed, noisy data

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_20.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.20.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    cox
    -0.27
    åĮºåĿĹ
    -0.27
    æŀģé«ĺ
    -0.26
    ifen
    -0.26
    访
    -0.26
     Frances
    -0.26
    .isEmpty
    -0.25
    -scroll
    -0.24
    uest
    -0.24
    ufe
    -0.24
    POSITIVE LOGITS
     slow
    0.28
    alion
    0.27
     Pick
    0.27
    -equ
    0.27
     pick
    0.27
     slowly
    0.26
    èĩ
    0.26
    æħ¢
    0.25
    ADVERTISEMENT
    0.25
    &t
    0.25
    Activations Density 6.173%

    No Known Activations