© 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. 21-GEMMASCOPE-TRANSCODER-16K
    7. 13350
    Prev
    Next
    INDEX
    Explanations

    mentions of sports games and results, especially words like "score" and "run", and also seems to trigger on descriptions of tests/experiments, math/numbers and study evaluations

    oai_token-act-pair · gemini-2.0-flash

    the number "two"

    np_token-act-pair-logits · gpt-4o-mini

    Number two

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.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
     ten
    -1.14
     eleven
    -1.10
     Eleven
    -0.99
    Ten
    -0.99
     nine
    -0.98
     eight
    -0.96
     twelve
    -0.94
     Ten
    -0.94
     thirteen
    -0.93
    Eleven
    -0.93
    POSITIVE LOGITS
     two
    1.97
    two
    1.77
    Two
    1.63
     Two
    1.60
     deux
    1.52
     TWO
    1.49
    TWO
    1.45
     zwei
    1.30
     två
    1.28
     dois
    1.27
    Activations Density 3.798%

    No Known Activations