© 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. Gemma-2-2B
    5. 22-GEMMASCOPE-TRANSCODER-16K
    7. 4607
    Prev
    Next
    INDEX
    Explanations

    sentences where the author will now point something out.

    oai_token-act-pair · gemini-2.0-flash

    are

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_22/width_16k/average_l0_15
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     purpoſe
    -0.77
     ―――――
    -0.76
    #+#
    -0.74
     cdti
    -0.73
    makeText
    -0.72
     ་་
    -0.70
     crdi
    -0.70
     ♂
    -0.69
     diaphragm
    -0.68
     Efq
    -0.68
    POSITIVE LOGITS
     four
    0.94
     five
    0.93
     the
    0.93
     three
    0.88
     six
    0.88
     eight
    0.86
     seven
    0.79
     a
    0.78
     nine
    0.73
     my
    0.71
    Activations Density 5.755%

    No Known Activations