© 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. 30-TRANSCODER-HP
    7. 95759
    Prev
    Next
    INDEX
    Explanations

    code documentation

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     trab
    -0.29
    ryn
    -0.28
     Breed
    -0.28
     perch
    -0.27
    ARK
    -0.26
     squir
    -0.25
     breeding
    -0.25
    çĮ¬
    -0.25
    ICH
    -0.25
     reacting
    -0.25
    POSITIVE LOGITS
     covering
    0.27
    第ä¸Ģ次
    0.27
    ene
    0.25
    covering
    0.25
    domains
    0.24
    hits
    0.24
    çĤĬ
    0.24
    æĵĺ
    0.24
    swire
    0.24
    è¦ĨçĽĸ
    0.24
    Activations Density 0.882%

    No Known Activations