© 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. 26003
    Prev
    Next
    INDEX
    Explanations

    United States politics

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

    The neuron fires strongly on tokens that contain non-letter characters (e.g. digits, apostrophes, punctuation), such as contractions, numbers, and abbreviations.

    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
    ׇ
    -0.96
    umque
    -0.93
    ￠
    -0.91
    :^(
    -0.91
     '--
    -0.89
    ὸν
    -0.89
     their
    -0.87
    irnov
    -0.86
    歳の
    -0.84
     taurus
    -0.84
    POSITIVE LOGITS
     бампер
    1.38
    ּוֹ
    1.30
    ִּ
    1.14
     yyb
    0.97
    ֶּ
    0.96
    чего
    0.90
     Pozna
    0.89
     czarny
    0.89
    kenstock
    0.89
     unn
    0.89
    Activations Density 0.050%

    No Known Activations