© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 27761
    Prev
    Next
    INDEX
    Explanations

    game mechanics and programming concepts

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

    The neuron is detecting programming‐style identifiers and constants (uppercase or mixed‐case code tokens and macro/struct names) within source 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_22/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
     on
    -2.11
     in
    -1.86
     more
    -1.70
     is
    -1.66
     with
    -1.59
     all
    -1.59
     now
    -1.46
     mainly
    -1.46
     first
    -1.43
     no
    -1.41
    POSITIVE LOGITS
     dénon
    2.02
    معلومات
    1.78
     colorear
    1.73
     obé
    1.69
     relâche
    1.66
    متن
    1.66
     chyb
    1.65
     jednod
    1.64
    Lø
    1.60
     tyto
    1.59
    Activations Density 0.104%

    No Known Activations