INDEX

Explanations

to

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

 reflected

-0.33

 reflecting

-0.30

Reflect

-0.28

 reflective

-0.28

 reflections

-0.28

 reflects

-0.28

reflect

-0.28

Ã³s

-0.27

 reflection

-0.27

 Reflect

-0.26

POSITIVE LOGITS

 discover

0.40

 discovery

0.35

 find

0.34

go

0.33

 read

0.30

åıĳçİ°

0.30

 goto

0.30

 Discovery

0.29

 audition

0.28

ä¸įåģľçļĦ

0.28

Activations Density 0.071%