© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 39039
    Prev
    Next
    INDEX
    Explanations

    programming declarations

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron fires on Java type or constant identifiers (tokens beginning with an uppercase letter, e.g. class names or enum constants).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     after
    -1.97
     every
    -1.82
     if
    -1.78
     through
    -1.75
    Those
    -1.66
    These
    -1.62
     took
    -1.61
     unrelenting
    -1.60
    Through
    -1.55
     give
    -1.54
    POSITIVE LOGITS
     amazingly
    1.83
     unbelievably
    1.65
     ridiculously
    1.63
     nonchal
    1.62
     fordi
    1.62
    嗬
    1.56
    -
    1.56
     kinda
    1.56
     notwendig
    1.55
     soooo
    1.55
    Activations Density 0.037%

    No Known Activations