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

    Neuronpedia

    INDEX
    Explanations

    defining prohibited content

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     flavoring
    0.35
    洄
    0.34
     stabbing
    0.33
     eliminación
    0.32
     spilling
    0.32
    กิจกรรม
    0.32
     secretion
    0.32
     wrapping
    0.32
     holen
    0.32
    रत
    0.31
    POSITIVE LOGITS
     could
    0.50
     would
    0.45
     relates
    0.41
     are
    0.40
     might
    0.39
     pourrait
    0.38
    could
    0.38
     COULD
    0.35
     enhances
    0.35
     છે
    0.34
    Act Density 0.031%

    No Known Activations