INDEX

Explanations

well

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

Ð¸Ð·Ð°ÑĤÐ¾ÑĢ

-0.31

edly

-0.26

atile

-0.26

 sided

-0.26

 natural

-0.26

ropsych

-0.25

mine

-0.25

Ð°ÑĢÐ°ÑĤ

-0.25

alytics

-0.25

 arty

-0.25

POSITIVE LOGITS

ijn

0.32

 well

0.31

åħ¬å¹³

0.30

çĨŁæĤī

0.29

æľªçŁ¥

0.27

 known

0.27

 Known

0.27

éĻĮçĶŁ

0.26

 Unknown

0.26

well

0.26

Activations Density 0.010%