INDEX

Explanations

return or print statements

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

 исчез

-0.87

 remplacer

-0.84

 terminer

-0.84

absence

-0.83

ofing

-0.82

 utiliser

-0.81

 chercher

-0.78

 veränder

-0.78

如果没有

-0.78

更是

-0.77

POSITIVE LOGITS

 failed

1.38

re

1.12

 remain

1.10

 retry

1.08

 redo

1.06

 stay

1.05

 repeat

1.04

 back

1.04

重新

1.04

 again

1.03

Activations Density 0.004%