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

éķ¿çŁŃ

-0.27

usher

-0.26

ennie

-0.26

æĿ¾å¼Ľ

-0.26

ross

-0.26

 deformation

-0.25

 stretches

-0.25

Mus

-0.25

standing

-0.24

Invariant

-0.24

POSITIVE LOGITS

 Normal

0.31

Normal

0.28

incipal

0.28

arine

0.28

>Main

0.27

assium

0.27

çĶŁçĲĨ

0.26

amodel

0.25

çĽĽ

0.25

EQ

0.25

Activations Density 0.886%