© 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. 19392
    Prev
    Next
    INDEX
    Explanations

    uncertain research findings

    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
    ÑģÑĤв
    -0.32
    igh
    -0.31
    ç¥ĸ
    -0.28
    KL
    -0.27
    ç´§
    -0.27
     stad
    -0.26
    -script
    -0.26
    wald
    -0.26
    ough
    -0.26
    R
    -0.26
    POSITIVE LOGITS
    鼾
    0.30
    Mixed
    0.28
    undef
    0.27
     bölümü
    0.27
    缺ä¹ı
    0.27
    Unavailable
    0.26
    éĹ®é¢ĺæĺ¯
    0.26
     mixed
    0.26
     unavailable
    0.26
    unsupported
    0.26
    Activations Density 1.305%

    No Known Activations