© 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. 21311
    Prev
    Next
    INDEX
    Explanations

    conditional checks and data descriptions

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

    The neuron fires on code or template identifiers—especially those starting with an uppercase letter (e.g. type or variable names like “Data,” “Collection,” “IsCollection,” “link,” “section,” “url,” etc.).

    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
     leaſt
    -1.35
     siste
    -1.18
    لی
    -1.16
    itati
    -1.15
    gewä
    -1.14
     تعدى
    -1.12
    狁
    -1.12
    ába
    -1.12
    ѡ
    -1.11
    тех
    -1.10
    POSITIVE LOGITS
     only
    2.44
     more
    1.74
     merely
    1.64
     just
    1.59
     that
    1.59
     próprio
    1.48
     because
    1.41
     other
    1.40
     jedynie
    1.38
     smaller
    1.37
    Activations Density 0.384%

    No Known Activations