INDEX

Explanations

patent figure descriptions

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

başı

-0.79

kops

-0.78

اهش

-0.76

 artikkeli

-0.75

Someone

-0.75

tır

-0.74

FileUtil

-0.73

kusi

-0.73

চ্

-0.73

 Также

-0.72

POSITIVE LOGITS

Fig

1.42

FIG

1.36

Fig

1.30

FIG

1.21

Figs

1.04

fig

1.00

 Patent

0.93

Рис

0.86

了过来

0.85

Figure

0.84

Activations Density 0.039%