© 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. 22-TRANSCODER-HP
    7. 98720
    Prev
    Next
    INDEX
    Explanations

    Article start

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_22.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.22.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    çļĦ巨大
    -0.26
    (Return
    -0.25
    ãģ§ãģĻãģĹ
    -0.25
    â¾
    -0.25
    entic
    -0.25
    ãģ³
    -0.24
    迨
    -0.24
    ä¸Ģæĸ¹éĿ¢
    -0.24
     nonlinear
    -0.24
    Å¡t
    -0.24
    POSITIVE LOGITS
    rench
    0.27
    SURE
    0.26
     Mounted
    0.26
    åĨįè§ģ
    0.26
     spender
    0.25
    phas
    0.24
    ERSION
    0.24
    mounted
    0.24
    åĨį次
    0.24
    utable
    0.24
    Activations Density 0.007%

    No Known Activations