© 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. 19158
    Prev
    Next
    INDEX
    Explanations

    start

    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
    çĩ®
    -0.30
    pii
    -0.28
    isclosed
    -0.27
    认åIJĮ
    -0.26
    åŁĥå°Ķ
    -0.25
    _SYN
    -0.25
    roid
    -0.25
     lã
    -0.24
    éĻĦ
    -0.24
    acks
    -0.24
    POSITIVE LOGITS
     start
    0.56
    start
    0.52
     starts
    0.47
     Start
    0.46
    èµ·
    0.46
     starting
    0.44
    /start
    0.44
    Start
    0.43
    .start
    0.43
    å§ĭ
    0.43
    Activations Density 0.276%

    No Known Activations