© 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. 33-TRANSCODER-HP
    7. 53640
    Prev
    Next
    INDEX
    Explanations

    of

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_33.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.33.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    onical
    -0.33
    éĩį
    -0.29
    å®ģ
    -0.29
    ettel
    -0.25
    BY
    -0.25
    REAM
    -0.25
    onders
    -0.24
    оде
    -0.24
    åħ¥ãĤĬ
    -0.23
    $',
    -0.23
    POSITIVE LOGITS
     of
    0.47
    _Of
    0.30
    	of
    0.30
    类似çļĦ
    0.28
     của
    0.28
    _of
    0.28
    åĢĴåľ¨
    0.28
     in
    0.28
    OfFile
    0.27
    OfType
    0.27
    Activations Density 0.007%

    No Known Activations