© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 38583
    Prev
    Next
    INDEX
    Explanations

    verbs denoting actions

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

    The neuron is especially sensitive to dynamic action verbs in their “-ing” (present‐participle/gerund) form that describe performing or offering activities.

    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
     need
    -1.05
     attempt
    -1.03
    还想
    -1.02
     instead
    -1.02
     one
    -1.01
    等級
    -0.99
     alguns
    -0.99
     for
    -0.98
    ַּ
    -0.98
    了一些
    -0.97
    POSITIVE LOGITS
     azonban
    1.52
     people
    1.24
     kuitenkin
    1.20
    jih
    1.19
    Organizador
    1.17
     especially
    1.16
     však
    1.14
    especially
    1.12
     espes
    1.11
    Especially
    1.09
    Activations Density 0.041%

    No Known Activations