© 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. 19279
    Prev
    Next
    INDEX
    Explanations

    website sections

    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
     Quant
    -0.27
    ILT
    -0.27
    irts
    -0.27
    uars
    -0.26
    Quant
    -0.26
    ITAL
    -0.26
    Que
    -0.26
    gles
    -0.26
     Vulcan
    -0.25
    çļĦå±ĢéĿ¢
    -0.25
    POSITIVE LOGITS
     bar
    0.46
     box
    0.40
    bar
    0.40
     area
    0.39
     label
    0.38
    rame
    0.37
     comment
    0.37
     line
    0.37
    able
    0.34
     statement
    0.34
    Activations Density 1.497%

    No Known Activations