© 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. 19419
    Prev
    Next
    INDEX
    Explanations

    Rules, systems, and safety

    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
    mates
    -0.29
     Wh
    -0.27
    means
    -0.27
     means
    -0.27
     wh
    -0.26
    abee
    -0.26
    OTT
    -0.26
    rette
    -0.26
     sue
    -0.25
     Threat
    -0.25
    POSITIVE LOGITS
    aravel
    0.27
     polar
    0.27
    Repository
    0.26
     Polar
    0.25
    icum
    0.25
     polarization
    0.25
    éĢļ
    0.24
    è·Ľ
    0.24
    hib
    0.24
    pure
    0.24
    Activations Density 0.089%

    No Known Activations