INDEX
Explanations
Polish and Russian math and sayings
New Auto-Interp
Negative Logits
connaît
0.65
superfici
0.57
Apesar
0.55
mógł
0.54
cams
0.54
Несмотря
0.54
oamenii
0.54
numerosos
0.54
podría
0.52
swore
0.52
POSITIVE LOGITS
displaystyle
0.71
displaystyle
0.68
}^{0.66
rm
0.61
Displaystyle
0.60
mathbf
0.59
mathbb
0.58
bf
0.58
leq
0.58
}^{+}0.56
Activations Density 0.005%