INDEX

Explanations

responsible for tasks

sequences following specific words

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

Lever

0.45

aspid

0.44

 втра

0.44

LIVE

0.43

ogna

0.43

 Lever

0.42

""".

0.42

lever

0.41

":[{"

0.41

chemistry

0.41

POSITIVE LOGITS

 bezahlen

0.48

ों

0.48

 выбо

0.46

 ranchers

0.46

얀

0.45

 anglers

0.44

 プレゼント

0.43

<0x80>

0.43

మ్మ

0.43

Activations Density 0.001%