© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 128826
    Prev
    Next
    INDEX
    Explanations

    Code and programming constructs

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

    The neuron fires on subword fragments of programming identifiers (e.g. “Item,” “ctor,” “of”) in code.

    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
     Даже
    -1.73
     Without
    -1.73
     Despite
    -1.69
     Füßen
    -1.67
     Using
    -1.66
    Only
    -1.66
    Sometimes
    -1.66
     This
    -1.65
     Sometimes
    -1.65
     understand
    -1.62
    POSITIVE LOGITS
    艽
    1.69
    ⸥
    1.51
    complish
    1.49
    Botswana
    1.49
    weakSelf
    1.48
     لعبة
    1.43
    Zambia
    1.42
     noe
    1.41
        
    1.40
     avem
    1.39
    Activations Density 0.000%

    No Known Activations