INDEX

Explanations

what makes

what makes a

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 without

0.46

 без

0.45

 tâche

0.43

 Without

0.43

 用于

0.42

 αφού

0.41

 όπου

0.41

有助于

0.40

 посредством

0.40

デ

0.39

POSITIVE LOGITS

 qualifies

0.48

ells

0.45

tick

0.44

aroos

0.42

 decides

0.42

appt

0.42

successful

0.42

mers

0.41

trigger

0.41

these

0.41

Activations Density 0.043%