INDEX

Explanations

of

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-ring

-0.30

zy

-0.29

Prov

-0.26

çº¨

-0.26

-chat

-0.26

ãģĽãĤĭ

-0.26

gens

-0.24

gia

-0.24

pecting

-0.24

unded

-0.24

POSITIVE LOGITS

0.28

 Tools

0.27

 aprox

0.26

 approximately

0.26

çº¦

0.26

çº¦åĲĪ

0.25

 reach

0.25

à¹Ģà¸Ħà¸¢

0.24

 Raised

0.23

 Voor

0.23

Activations Density 0.019%