© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 66958
    Prev
    Next
    INDEX
    Explanations

    starts of questions or code

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

    The neuron activates on programming–style tokens (e.g. keywords like “defer,” “func,” “if” or Camel-cased method names like “Dispose,” “SetOutput”) and thus appears to detect code fragments.

    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
     kép
    -0.96
     my
    -0.93
     their
    -0.93
    #@
    -0.84
    uming
    -0.84
     fresca
    -0.84
     you
    -0.84
     for
    -0.84
     DNS
    -0.84
    Eksteraj
    -0.83
    POSITIVE LOGITS
     alimentazione
    0.99
    Détails
    0.97
     tratan
    0.96
    ungkinkan
    0.90
     orch
    0.89
     الحكومة
    0.89
    Irlande
    0.88
     chén
    0.88
     conic
    0.88
     السياس
    0.87
    Activations Density 0.000%

    No Known Activations