© 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. 13017
    Prev
    Next
    INDEX
    Explanations

    constructor

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

    The neuron is tuned to pick out numeric tokens—numbers in comments, code (line numbers, literal values), and similar numeric references.

    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
     proguardFiles
    0.63
     "_
    0.61
     tweaked
    0.61
     WINTER
    0.61
     tendrá
    0.59
     androidTest
    0.59
     ____
    0.57
     ___
    0.57
     realms
    0.57
     //}
    0.57
    POSITIVE LOGITS
    constructor
    1.13
     constructor
    0.96
    constru
    0.77
    Constructor
    0.76
    construction
    0.72
    构造
    0.69
     Constructor
    0.66
     construction
    0.64
     конструк
    0.63
    constr
    0.63
    Activations Density 0.003%

    No Known Activations