© 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. 67606
    Prev
    Next
    INDEX
    Explanations

    introduces clause

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

    This neuron activates on interrogative or demonstrative function words (like “что,” “какие,” “diese,” etc.) that commonly appear in questions.

    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
     on
    -2.03
     or
    -2.02
    e
    -1.98
    r
    -1.95
    We
    -1.84
    f
    -1.73
     possibly
    -1.58
    la
    -1.57
    我们
    -1.57
     mainly
    -1.55
    POSITIVE LOGITS
    despite
    1.88
    神情
    1.86
    міні
    1.83
     aliados
    1.79
    ፨
    1.67
     célèbre
    1.64
    棡
    1.63
     ЛИ
    1.59
     accompagné
    1.58
     genellikle
    1.58
    Activations Density 0.017%

    No Known Activations