© 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. 7630
    Prev
    Next
    INDEX
    Explanations

    keep getting

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

    The neuron fires on occurrences of the verb “keep” (and its morphological variants like “kept”).

    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
    ↵↵
    -3.00
    .
    -2.89
    "
    -2.86
    𓋼
    -2.72
    s
    -2.70
    e
    -2.67
    ,"
    -2.55
    )
    -2.53
     '
    -2.42
    <
    -2.41
    POSITIVE LOGITS
    袛
    2.83
    幷
    2.72
    ﱠ
    2.69
    𝝈
    2.63
    >();
    2.61
    ꙁ
    2.59
     Гі
    2.58
    ：
    2.58
    ());
    2.56
    нутри
    2.55
    Activations Density 0.020%

    No Known Activations