© 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. 128494
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on language‐label annotations (e.g. the language name plus colon in parenthetical translations like “French:”, “German:”, “Portuguese:”, etc.).

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

    foreign language names

    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
     niektó
    -1.10
     nuestra
    -1.03
    これに
    -1.01
    それに
    -0.99
     verksam
    -0.97
     några
    -0.96
     algunos
    -0.96
    こういった
    -0.96
     these
    -0.95
     These
    -0.94
    POSITIVE LOGITS
     белый
    1.06
     آمریکا
    1.03
    <!--
    1.02
     Goi
    1.02
     laik
    1.02
    organised
    1.00
    bär
    0.99
    LOYD
    0.99
     baut
    0.99
     milion
    0.98
    Activations Density 0.012%

    No Known Activations