© 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. 18621
    Prev
    Next
    INDEX
    Explanations

    Clothing, specifically coats

    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
    人类
    -0.39
     human
    -0.35
    -human
    -0.34
    human
    -0.31
    _human
    -0.30
    人é¡ŀ
    -0.29
     humans
    -0.29
     Human
    -0.28
    Human
    -0.27
    ndon
    -0.27
    POSITIVE LOGITS
    å¼ıçļĦ
    0.28
    å¼ı
    0.27
    material
    0.27
     material
    0.27
    衬
    0.26
    æļĹæļĹ
    0.26
     smooth
    0.26
    smooth
    0.26
     thick
    0.25
    æĢ¼
    0.25
    Activations Density 0.190%

    No Known Activations