INDEX

Explanations

defined as

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mata

-0.08

aux

-0.07

Ã½t

-0.07

=>"

-0.07

parator

-0.07

hangi

-0.07

orida

-0.07

è²ī

-0.07

æīįèĥ½æľī

-0.07

TECT

-0.06

POSITIVE LOGITS

 combinations

0.08

 simply

0.08

sum

0.08

 depends

0.08

åıĸåĨ³äºİ

0.07

sum

0.07

 expressions

0.07

ä¹ĺ

0.07

à¹Ħà¸Ķà¹īà¸Īà¸²à¸ģ

0.07

 iter

0.07

Activations Density 0.026%