© 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. 10-GEMMASCOPE-RES-131K
    7. 59579
    Prev
    Next
    INDEX
    Explanations

    php code snippets

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

    This neuron fires primarily on numeric literals (and similar symbol‐only tokens), i.e. standalone numbers and punctuation‐style tokens rather than words.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     hés
    -1.66
     porque
    -1.63
     esimer
    -1.61
     daje
    -1.59
     BECAUSE
    -1.56
     ajouter
    -1.55
     магазин
    -1.54
     baterie
    -1.52
     pontos
    -1.51
     rése
    -1.49
    POSITIVE LOGITS
     Regardless
    1.53
     in
    1.48
     Despite
    1.46
     They
    1.39
     Although
    1.37
     by
    1.35
     With
    1.35
    бю
    1.34
     at
    1.33
    ክ
    1.28
    Activations Density 0.018%

    No Known Activations