© 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. 127648
    Prev
    Next
    INDEX
    Explanations

    The neuron consistently lights up on “weird” non‐sentence tokens—things like ellipses (“…”), code symbols (the “=>” arrow), standalone uppercase or punctuation‐heavy tokens, etc.

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408

    programming keywords

    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
    Negative Logits
    ֩
    -1.14
     répondu
    -1.04
    |}{\
    -1.02
     Horizonte
    -1.01
    ક્
    -1.00
    Procedimento
    -1.00
    umán
    -0.99
     There
    -0.97
    许多
    -0.97
    ագրություններ
    -0.97
    POSITIVE LOGITS
     Août
    1.20
     Juillet
    1.16
     Février
    1.11
     Décembre
    0.96
     kompet
    0.95
     Espagne
    0.95
     Să
    0.94
     DÉ
    0.94
     считают
    0.93
     NORTH
    0.93
    Activations Density 0.020%

    No Known Activations