© 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. 7983
    Prev
    Next
    INDEX
    Explanations

    multilingual beginnings

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

    The neuron sharply fires on any token containing the three‐letter sequence “val” (or its Cyrillic counterpart “вал”), picking up stems like “valid,” “Valdemar,” “пробовал,” 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
    ㈮
    -2.94
    庒
    -2.89
    儅
    -2.89
    凇
    -2.80
    阋
    -2.73
    侂
    -2.73
    鬭
    -2.69
     xxiii
    -2.61
    夘
    -2.59
     ソックス
    -2.58
    POSITIVE LOGITS
    e
    2.75
    N
    2.72
     was
    2.64
    To
    2.44
     más
    2.41
    </strong>
    2.38
     
    2.31
    してます
    2.17
     stuffy
    2.16
     الأربع
    2.16
    Activations Density 0.001%

    No Known Activations