INDEX

Explanations

code syntax punctuation

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

and

-1.91

其他的

-1.39

 différents

-1.34

神情

-1.29

 velmi

-1.27

 andere

-1.27

 două

-1.27

,;

-1.23

 faciliter

-1.23

 dejamos

-1.21

POSITIVE LOGITS

 despite

1.66

</strong>

1.62

</h2>

1.48

 Usually

1.44

</th>

1.41

 seemed

1.40

 Being

1.34

__":

1.34

 without

1.32

 This

1.29

Activations Density 0.077%