© 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. 33755
    Prev
    Next
    INDEX
    Explanations

    zygGraph start tokens

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

    The neuron fires strongly on lines that introduce or delimit code blocks—e.g. opening braces and keywords like else, try, and catch.

    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
     was
    -2.33
    providedIn
    -2.25
     is
    -2.09
     two
    -1.77
     has
    -1.71
     All
    -1.70
     
    -1.68
     (
    -1.66
     A
    -1.64
     so
    -1.60
    POSITIVE LOGITS
    mittelt
    2.02
    ナイキ
    1.99
     wichtige
    1.99
    饪
    1.97
    蠵
    1.94
    瑧
    1.92
     productividad
    1.88
    憮
    1.82
    慳
    1.82
    
    1.82
    Activations Density 0.002%

    No Known Activations