© 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. 6192
    Prev
    Next
    INDEX
    Explanations

    public class declarations

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

    The neuron specializes in spotting small function words that introduce relationships or negations—e.g. tokens like “is not,” “of,” and “with.”

    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
     (@
    0.63
     தம்
    0.62
     @
    0.61
     லட்ச
    0.58
     paltry
    0.58
    @
    0.58
     நிறைவே
    0.58
     Staffordshire
    0.56
    CollectedField
    0.56
     TreeMap
    0.55
    POSITIVE LOGITS
    Internal
    0.65
     [
    0.62
    ///
    0.59
    [
    0.59
     \[
    0.58
    ickt
    0.55
    \[
    0.55
    internal
    0.54
    legacy
    0.54
    ()[
    0.54
    Activations Density 0.003%

    No Known Activations