© 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. 22-GEMMASCOPE-RES-131K
    7. 26671
    Prev
    Next
    INDEX
    Explanations

    parenthetical annotations and corrections

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

    The neuron detects editorial markup tokens (e.g. “[sic]”, “(cq)”, “corrected”, “(sp)”, “™”, italic/formatting tags, 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_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
    蛱
    -1.22
    dza
    -1.13
    XmlAttribute
    -1.10
    cuál
    -1.09
    phòng
    -1.09
    )।
    -1.09
     textos
    -1.09
    úsculo
    -1.09
    踱
    -1.08
    趼
    -1.08
    POSITIVE LOGITS
    )
    1.32
    .,
    1.27
    "
    1.23
    sic
    1.07
     acompañar
    1.05
    –
    1.05
    biť
    1.04
     acrescent
    1.02
    n
    1.02
    '
    1.01
    Activations Density 0.003%

    No Known Activations