INDEX

Explanations

to + verb

to be helpful

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

nah

0.46

 inap

0.45

 முயற்சிய

0.44

 تح

0.44

 când

0.43

."))

0.43

 Sharma

0.42

អំ

0.42

 اقد

0.42

 überhaupt

0.42

POSITIVE LOGITS

 ώστε

0.76

 بحيث

0.68

 যাতে

0.67

 zodat

0.64

 ताकि

0.61

 withstand

0.60

 максимально

0.56

以便

0.53

 maximise

0.52

เพื่อให้

0.52

Activations Density 0.197%