© 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. 18460
    Prev
    Next
    INDEX
    Explanations

    Sand

    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
    ENCE
    -0.27
    ä»İ严治
    -0.26
    clair
    -0.25
    æł¸å¿ĥ
    -0.25
    éĶ»çĤ¼
    -0.25
    å¼ĢæĿ¥
    -0.25
    ä¸įä½ı
    -0.25
    _splits
    -0.25
     Muse
    -0.24
    ocate
    -0.24
    POSITIVE LOGITS
     Sand
    0.45
    Sand
    0.44
     Oak
    0.42
     Los
    0.38
     Arg
    0.38
    Los
    0.35
     SAND
    0.35
     Lawrence
    0.35
    Oak
    0.34
    sand
    0.34
    Activations Density 9.493%

    No Known Activations