© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 25242
    Prev
    Next
    INDEX
    Explanations

    code blocks and functions

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

    This neuron detects programming code segments by firing on code‐related syntax tokens (indentation, punctuation like “:”, “def”, underscores in identifiers, function names such as “send,” etc.).

    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
    やっぱ
    -0.92
    頹
    -0.90
     rése
    -0.89
     подска
    -0.88
    develop
    -0.87
    のかな
    -0.86
    tritte
    -0.85
    Final
    -0.85
    finally
    -0.85
     revenir
    -0.82
    POSITIVE LOGITS
     postponement
    1.02
     commencement
    1.00
    เริ่ม
    0.98
     promissory
    0.97
    等待
    0.97
     waiting
    0.95
     temporarily
    0.95
     pending
    0.93
     started
    0.92
     leaving
    0.91
    Activations Density 0.035%

    No Known Activations