© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 103362
    Prev
    Next
    INDEX
    Explanations

    action followed by destination

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

    The neuron fires on action words—i.e. verbs or command‐style tokens (like “GoTo,” “addto,” “StringTo,” Chinese verbs such as “到,” “查看,” “拖,” 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
    蠖
    -2.78
    astify
    -2.30
    "]);
    -2.23
    颙
    -2.23
    "]));
    -2.22
    鳊
    -2.19
    </i>
    -2.17
    》。
    -2.17
    ンの
    -2.16
    ：　
    -2.14
    POSITIVE LOGITS
    ”
    3.02
     deine
    2.53
    在他
    2.36
     elevate
    2.36
     jsou
    2.31
    您的
    2.30
    -
    2.27
     have
    2.25
     ‘
    2.22
     werden
    2.17
    Activations Density 0.001%

    No Known Activations