© 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. 22234
    Prev
    Next
    INDEX
    Explanations

    could be measured/demonstrated

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

    This neuron detects occurrences of the modal auxiliary “could” (especially when followed by “be” in passive‐voice constructions).

    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
    ;%%
    -1.75
    "
    -1.33
    
    -1.21
     "../../../../
    -1.20
    +)
    -1.15
    ёв
    -1.13
    reportWebVitals
    -1.10
    ed
    -1.10
    ੇ
    -1.09
     a
    -1.08
    POSITIVE LOGITS
    הגב
    1.28
    Примі
    1.23
    </em>
    1.23
     удалось
    1.23
     получилось
    1.22
    Переваги
    1.21
    Familia
    1.20
     funcionan
    1.18
     INSTIT
    1.17
     devido
    1.17
    Activations Density 0.055%

    No Known Activations