© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 2980
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on code‐like or API/documentation tokens (e.g. type names, identifiers, operators)—in other words, it’s keyed to programming–style vocabulary.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     semidefinite
    0.64
     conic
    0.63
    产业
    0.57
     hillside
    0.57
    ∬
    0.57
     कार्तिक
    0.56
    🚴
    0.56
     landslide
    0.55
    蝼
    0.55
     sevent
    0.54
    POSITIVE LOGITS
     String
    2.03
    字符串
    2.03
     string
    2.00
    String
    1.95
    文字列
    1.95
    string
    1.94
     strings
    1.88
     Strings
    1.79
    STRING
    1.75
    strings
    1.71
    Activations Density 0.140%

    No Known Activations