© 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. 33826
    Prev
    Next
    INDEX
    Explanations

    proper names and associated terms

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

    The neuron fires on capitalized tokens—particularly proper names and technical or titular terms.

    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
    buffer
    -0.94
     branco
    -0.90
    comme
    -0.86
     гостиной
    -0.83
    parece
    -0.83
     Clinton
    -0.82
     abundance
    -0.82
    मे
    -0.82
     domicili
    -0.80
    accumulator
    -0.80
    POSITIVE LOGITS
     thereof
    0.90
    numel
    0.90
    などなど
    0.89
     petugas
    0.88
     there
    0.87
     niitä
    0.84
    ยะ
    0.82
     '*':
    0.81
    ("}");
    0.81
     JsonConvert
    0.79
    Activations Density 0.028%

    No Known Activations