Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    say verbs

    np_max-act-logits · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     дают
    0.86
     permettent
    0.79
     включают
    0.79
     էին
    0.77
     позволяют
    0.76
     proporcionan
    0.73
    Have
    0.73
     offrent
    0.72
     occur
    0.71
     ofrecen
    0.70
    POSITIVE LOGITS
     wants
    2.17
     knows
    2.16
     thinks
    2.03
     loves
    1.98
     understands
    1.97
     believes
    1.87
     хочет
    1.84
     prefers
    1.82
     expects
    1.78
     hates
    1.78
    Act Density 0.434%

    No Known Activations