© 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. 22-GEMMASCOPE-RES-131K
    7. 15693
    Prev
    Next
    INDEX
    Explanations

    code keywords and syntax

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

    The neuron is tuned to pick out programming‐language syntax elements—braces, keywords, and identifiers (function, type, and variable names) in 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
     Therefore
    -1.05
     Again
    -0.99
     again
    -0.96
    を与える
    -0.95
    ضات
    -0.93
     what
    -0.86
    нову
    -0.84
    bür
    -0.82
    ICLES
    -0.81
    Again
    -0.80
    POSITIVE LOGITS
    ängerin
    0.95
     edil
    0.91
    者に
    0.88
     meses
    0.88
    ūnas
    0.87
     宅
    0.85
    bitos
    0.84
    Ổ
    0.83
     Malaga
    0.83
     liberté
    0.83
    Activations Density 0.009%

    No Known Activations