INDEX

Explanations

numbered instructions

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

1.78

↵↵

1.32

↵↵↵

1.22

1.07

 also

1.06

<b>

1.05

 lieu

1.01

 strcmp

1.00

olipid

0.99

otoxin

0.98

POSITIVE LOGITS

الح

3.39

أس

3.34

ет

3.26

ве

3.24

склю

3.22

انس

3.20

ди

3.20

 Кор

3.19

чай

3.16

дня

3.15

Activations Density 0.004%