© 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. 127950
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on words that signal plurality or classification—especially terms like “types,” “different,” and other markers of multiple categories.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    types of

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    还不错
    -1.09
    Außerdem
    -1.05
    ４
    -1.02
    ３
    -0.99
    Bereich
    -0.99
    1
    -0.96
    అ
    -0.96
    Democratic
    -0.94
    データ
    -0.92
    jazdu
    -0.92
    POSITIVE LOGITS
     than
    1.41
     each
    1.23
     and
    1.20
     supposedly
    1.02
     pò
    1.00
     Había
    0.98
     might
    0.95
     bawat
    0.93
    お土産
    0.93
     trae
    0.93
    Activations Density 0.058%

    No Known Activations