© 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. 19182
    Prev
    Next
    INDEX
    Explanations

    technical issues

    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
    æĶ¶å½ķ
    -0.27
    anja
    -0.27
     outer
    -0.27
     inner
    -0.26
     act
    -0.26
    éĻIJå®ļ
    -0.26
    ox
    -0.25
    onia
    -0.25
     catch
    -0.25
    @endsection
    -0.25
    POSITIVE LOGITS
    ä¼ı
    0.30
    ulary
    0.30
    弧
    0.29
    çŁŃ
    0.28
    çħ¨
    0.28
    -hook
    0.27
     Shorts
    0.27
     hangs
    0.27
     hook
    0.27
     Arc
    0.27
    Activations Density 1.017%

    No Known Activations