INDEX

Explanations

for many years

especially often

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

whose

0.64

जिसे

0.60

完璧

0.57

nicht

0.52

 cuya

0.52

 інші

0.52

 אשר

0.52

which

0.51

לא

0.51

 slechts

0.51

POSITIVE LOGITS

 especially

0.58

 using

0.55

 espe

0.54

 both

0.53

 particularly

0.53

 kubectl

0.52

 özellikle

0.51

 github

0.50

经常

0.50

 gute

0.50

Activations Density 0.022%