© 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. 96427
    Prev
    Next
    INDEX
    Explanations

    things

    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
    åŃ¢
    -0.31
    who
    -0.28
    åĴĮåħ¶ä»ĸ
    -0.26
    åĽĽé¡¹
    -0.25
    -develop
    -0.25
     Appears
    -0.25
     boob
    -0.25
    åıĬåħ¶ä»ĸ
    -0.25
     versa
    -0.24
     who
    -0.24
    POSITIVE LOGITS
     wrong
    0.34
     that
    0.33
     matter
    0.29
    ãĤ½ãĥ³
    0.29
     we
    0.29
     vou
    0.28
     you
    0.28
    wrong
    0.27
    人们
    0.27
     learned
    0.27
    Activations Density 0.008%

    No Known Activations