INDEX

Explanations

functional outcomes

providing experimental outcomes

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

啊

0.45

でも

0.42

 वगैरह

0.42

 schoolchildren

0.41

之类的

0.41

 whatnot

0.40

哈哈

0.39

<unused2204>

0.39

봅시다

0.38

 でも

0.38

POSITIVE LOGITS

 allowing

1.38

 implying

1.30

 giving

1.22

 emphasizing

1.21

 ensuring

1.17

 suggesting

1.16

 providing

1.15

 making

1.14

 indicating

1.13

 enabling

1.13

Activations Density 1.038%