© 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. 113015
    Prev
    Next
    INDEX
    Explanations

    vehicular automatic

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

    The neuron fires strongly on “xe” tokens that are part of UTF-8 multibyte escape sequences (e.g. “\xe2\x80\x9c”), i.e. it detects Unicode-encoded characters.

    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
    鼱
    -2.61
    ”
    -2.48
    ’
    -2.39
     edif
    -2.33
    “
    -2.30
    葙
    -2.28
     they
    -2.23
    ]
    -2.22
     手绘
    -2.22
    袛
    -2.20
    POSITIVE LOGITS
     permiten
    2.84
    轱
    2.73
    </i>
    2.48
    NÍ
    2.48
     last
    2.44
    蕷
    2.42
     These
    2.31
     encuentre
    2.28
     แต่
    2.28
     same
    2.28
    Activations Density 0.005%

    No Known Activations