© 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. 31-TRANSCODER-HP
    7. 19185
    Prev
    Next
    INDEX
    Explanations

    Code, URLs

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     subject
    -0.29
    å¢ħ
    -0.28
     ITE
    -0.28
     closed
    -0.27
     ý
    -0.27
    æĹ·
    -0.26
    subject
    -0.25
     subjects
    -0.25
    Choices
    -0.25
    _choices
    -0.24
    POSITIVE LOGITS
    pons
    0.29
     Invest
    0.26
     hết
    0.26
    unks
    0.25
    uilder
    0.25
     Tank
    0.24
    undle
    0.24
     ún
    0.24
    å§ĭ
    0.24
    ÑģоÑĤ
    0.24
    Activations Density 0.061%

    No Known Activations