INDEX
    Explanations

    non-English languages

    The neuron fires on the Spanish word “Espero” (as in “Espero que esto te ayude”), i.e. spotting the polite “I hope this helps” opening in Spanish.

    New Auto-Interp
    Negative Logits
     ballot
    -0.08
     stamps
    -0.08
     Stamp
    -0.08
    ICH
    -0.07
     bureauc
    -0.07
     stamp
    -0.06
    Applications
    -0.06
    ich
    -0.06
    numpy
    -0.06
     Angle
    -0.06
    POSITIVE LOGITS
     espera
    0.08
    الم
    0.07
     Esper
    0.07
    0.07
     irre
    0.07
    ρέ
    0.07
     jsx
    0.06
    anim
    0.06
     glGet
    0.06
    0.06
    Act Density 0.004%

    No Known Activations