© 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. 33487
    Prev
    Next
    INDEX
    Explanations

    HTML template syntax

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

    This neuron activates on punctuation tokens that mark the ends of code blocks or statements—especially closing braces (“}”), closing brace–semicolon combinations (“};”), and similar delimiters.

    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
     also
    -1.05
     another
    -1.05
    性に
    -1.05
     to
    -0.98
     noiva
    -0.92
     acrí
    -0.92
    ּוֹ
    -0.91
    なのだ
    -0.89
     mangel
    -0.88
    することも
    -0.88
    POSITIVE LOGITS
    !")
    1.02
     OSX
    1.00
    だいぶ
    0.94
     osteo
    0.94
    registrer
    0.88
    彼が
    0.87
    他們
    0.87
    !')
    0.87
    生产
    0.86
     ساده
    0.86
    Activations Density 0.005%

    No Known Activations