© 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. 26018
    Prev
    Next
    INDEX
    Explanations

    be followed by there, sure, aware

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

    The neuron selectively fires on the standalone verb “be” (especially in its imperative or advisory uses).

    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
    quelize
    -2.61
    isième
    -2.16
    amente
    -2.09
    imetry
    -1.99
    izability
    -1.92
    nological
    -1.91
    aneity
    -1.91
    tifiable
    -1.88
    acterium
    -1.88
     desiring
    -1.88
    POSITIVE LOGITS
     able
    2.42
    為に
    2.22
     VERY
    1.98
     But
    1.97
     It
    1.96
    VIATIONS
    1.95
    *
    1.93
     verdaderamente
    1.87
    {
    1.87
     únicamente
    1.82
    Activations Density 0.017%

    No Known Activations