© 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. 95543
    Prev
    Next
    INDEX
    Explanations

    working

    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
     functioning
    -0.27
    eland
    -0.26
    __/
    -0.25
    æĭĮ
    -0.25
     fir
    -0.25
    滤
    -0.25
    earned
    -0.25
    olla
    -0.24
    ç»ıèIJ¥çIJĨ念
    -0.24
     exercising
    -0.24
    POSITIVE LOGITS
     beside
    0.35
     together
    0.31
     beyond
    0.31
     alongside
    0.28
    lemetry
    0.28
     вмеÑģÑĤе
    0.27
     behind
    0.27
     seven
    0.27
     side
    0.27
     adj
    0.27
    Activations Density 0.106%

    No Known Activations