INDEX

Explanations

describing high regard

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

"));

-2.47

 lidé

-2.42

釤

-2.39

</i>

-2.34

ᝢ

-2.31

ázaro

-2.28

ЗИ

-2.27

不明白

-2.16

 conspicuously

-2.11

撳

-2.11

POSITIVE LOGITS

</h4>

2.75

↵

2.44

 вони

2.19

’

2.08

ing

1.99

 間取り

1.97

ﱠ

1.94

’，

1.90

(?)

1.89

Activations Density 0.002%