INDEX

Explanations

the

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

 late

-0.39

æĻļæľŁ

-0.34

Later

-0.32

æ¯ıå¤©

-0.32

 later

-0.32

late

-0.31

Late

-0.31

later

-0.30

tar

-0.30

 Late

-0.29

POSITIVE LOGITS

è¦ĭ

0.25

.flink

0.25

abilidad

0.25

plets

0.24

ervo

0.24

Fizz

0.24

è¿ĩ

0.24

 spur

0.24

igu

0.24

ugal

0.24

Activations Density 0.103%