INDEX

Explanations

introducing examples and descriptions

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

<unused1778>

1.28

<unused524>

1.25

gpu

1.24

fw

1.24

嚨

1.22

숴

1.20

djang

1.19

 おい

1.19

}}^{*

1.19

<unused309>

1.19

POSITIVE LOGITS

 called

1.27

 Petra

1.27

El

1.24

Om

1.20

 Michael

1.17

 Benjamin

1.15

 named

1.14

 Maria

1.14

1.13

 Peter

1.11

Activations Density 1.004%