INDEX

Explanations

difficulties and absence

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

 նաև

-1.13

 动漫

-1.13

蒻

-1.12

今でも

-1.10

也知道

-1.09

割と

-1.09

ところに

-1.07

卡通

-1.05

ほら

-1.04

無事に

-1.02

POSITIVE LOGITS

 cannot

1.66

 only

1.62

 simply

1.53

 hardly

1.47

 δεν

1.43

ícias

1.36

 scarcely

1.35

 barely

1.33

ไม่

1.27

ไม่มี

1.25

Activations Density 0.144%