© 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. 96477
    Prev
    Next
    INDEX
    Explanations

    change or development

    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
    æĭ¿åĩºæĿ¥
    -0.32
    æݨåĩº
    -0.30
    æıIJåįĩ
    -0.28
     unlocks
    -0.28
     walmart
    -0.28
    baseline
    -0.26
     increasing
    -0.26
    åĬłå¿«å»ºè®¾
    -0.26
    åĩºä¾Ĩ
    -0.26
    iture
    -0.26
    POSITIVE LOGITS
     morph
    0.36
     hybrid
    0.29
     adapt
    0.29
    éĢļä¿Ĺ
    0.28
    ç²¾
    0.28
     gel
    0.28
     settled
    0.28
     fused
    0.27
     metam
    0.26
     was
    0.26
    Activations Density 0.103%

    No Known Activations