© 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. 10-GEMMASCOPE-RES-131K
    7. 84774
    Prev
    Next
    INDEX
    Explanations

    statements of expectation or obligation

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

    The neuron fires on occurrences of the modal verb “should” (often together with its following auxiliary like “be” or “have”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    iNdEx
    -1.69
     $-$
    -1.66
     eftersom
    -1.52
    with
    -1.52
     behandeln
    -1.49
    скільки
    -1.48
     Diputados
    -1.48
    lexus
    -1.48
     you
    -1.48
    可能な
    -1.48
    POSITIVE LOGITS
     normalement
    1.52
     should
    1.46
    should
    1.41
    s
    1.41
     ought
    1.40
    バイオ
    1.38
     resul
    1.36
     infal
    1.36
     înd
    1.34
    Probably
    1.34
    Activations Density 0.069%

    No Known Activations