© 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. 34-GEMMASCOPE-RES-131K
    7. 57664
    Prev
    Next
    INDEX
    Explanations

    can see or find

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

    This neuron fires on occurrences of the second‐person modal “can” (and related forms like “could”)—i.e. suggestions or permissions directed at “you.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    BorderFactory
    -1.20
     شروع
    -1.07
    SessionFactory
    -0.98
     ਵ
    -0.97
     mogą
    -0.96
     lustig
    -0.95
    assertAll
    -0.94
    Kut
    -0.92
    usaha
    -0.91
    QtGui
    -0.91
    POSITIVE LOGITS
     find
    1.70
     buy
    1.57
     see
    1.56
     get
    1.54
     expect
    1.41
     actually
    1.23
     usually
    1.22
     almost
    1.17
     hardly
    1.13
     hear
    1.13
    Activations Density 0.052%

    No Known Activations