© 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. 25199
    Prev
    Next
    INDEX
    Explanations

    code attribute contexts

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

    The neuron fires strongly on blocks of leading or inline whitespace (runs of spaces/tabs) that denote indentation in code or data.

    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
    นะนำ
    -1.09
    斡
    -0.85
     شلوار
    -0.84
    ずつ
    -0.84
     Dort
    -0.83
    したので
    -0.82
     which
    -0.81
    सन्दर्भ
    -0.80
    ilí
    -0.79
     Familienname
    -0.79
    POSITIVE LOGITS
     zwłaszcza
    0.96
    >[]
    0.96
    兲
    0.94
    大哥
    0.93
    textwidth
    0.93
    thas
    0.90
     Verbraucher
    0.90
     、
    0.90
    tees
    0.90
    นี้
    0.88
    Activations Density 0.062%

    No Known Activations