INDEX

Explanations

if, when, as, one, before

Frameworks

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

魉

0.38

তাসীন

0.36

 runes

0.35

रिडोर

0.34

рень

0.34

 alkene

0.34

ێنی

0.33

 تباہ

0.33

刄

0.33

जेंडर

0.32

POSITIVE LOGITS

If

0.45

The

0.45

While

0.45

As

0.43

One

0.41

Before

0.41

There

0.40

For

0.39

In

0.39

When

0.39

Activations Density 0.007%