INDEX

Explanations

a

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

atile

-0.27

å·¥ä½ľå²Ĺä½į

-0.27

.lv

-0.27

']}'

-0.26

 pubs

-0.25

 trophies

-0.25

_masks

-0.25

pany

-0.25

.masks

-0.25

 camer

-0.25

POSITIVE LOGITS

 science

0.30

 spine

0.30

 Science

0.30

fun

0.27

 spinner

0.27

æĢ§æĦŁ

0.26

 astrology

0.26

èĥİ

0.26

 animal

0.26

 guessing

0.26

Activations Density 0.884%