© 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. 34194
    Prev
    Next
    INDEX
    Explanations

    code structure definitions

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

    The neuron fires on the punctuation (parentheses and related symbols) that introduce parameter or type‐parameter lists in definitions (e.g. the “(” in class constructors, function signatures, annotations, and the “[”/“]” around generic parameters).

    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
     any
    -1.36
     there
    -1.32
     something
    -1.27
    نکته
    -1.25
     larger
    -1.21
     more
    -1.19
     another
    -1.16
     make
    -1.13
    فایل
    -1.12
    مرحله
    -1.11
    POSITIVE LOGITS
    भ
    1.09
    (
    1.07
     his
    1.07
     الإنسان
    1.06
    パソコン
    1.02
    /
    0.99
    unno
    0.98
    راه
    0.97
    oterapia
    0.97
     paradox
    0.96
    Activations Density 0.005%

    No Known Activations