INDEX

Explanations

recommendation or obligation

predicting should

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

 मजब

0.50

 приходится

0.43

ਸੀਂ

0.43

 ගැනීමට

0.43

υνα

0.41

 пришлось

0.41

 ಮಾಡಲು

0.40

 приходи

0.40

ഗണ

0.36

 مجبور

0.36

POSITIVE LOGITS

 should

4.75

should

4.63

 Should

4.44

Should

4.34

 SHOULD

4.16

应该

4.06

 powin

3.86

應該

3.84

 debería

3.83

ควร

3.83

Activations Density 0.776%