© 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. 96607
    Prev
    Next
    INDEX
    Explanations

    red

    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
     mo
    -0.27
    (fc
    -0.26
    огÑĢаÑĦ
    -0.26
     noct
    -0.24
    ngo
    -0.24
     pencils
    -0.24
     Friedman
    -0.24
    æ´»
    -0.23
     families
    -0.23
     tc
    -0.23
    POSITIVE LOGITS
     overst
    0.25
    erview
    0.25
    .increment
    0.25
    oa
    0.24
    汶
    0.24
    erville
    0.24
     Advis
    0.24
    \Common
    0.23
    ÃŃvel
    0.23
    æĪijçľĭ
    0.23
    Activations Density 0.099%

    No Known Activations