© 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. 128093
    Prev
    Next
    INDEX
    Explanations

    This neuron consistently lights up on words appearing at the start of sentences (i.e. initial capitalized tokens).

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

    consequence or result

    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
    '/',
    -0.96
     каждом
    -0.94
    paravant
    -0.93
    ssaint
    -0.89
    its
    -0.88
    chuan
    -0.87
     greifen
    -0.86
    égories
    -0.85
    cryption
    -0.85
    don
    -0.85
    POSITIVE LOGITS
     then
    2.38
     thus
    2.03
     which
    1.92
    sequently
    1.52
    Then
    1.48
     Thus
    1.48
     therefore
    1.37
     consequently
    1.36
     então
    1.35
    Thus
    1.34
    Activations Density 0.050%

    No Known Activations