INDEX

Explanations

subtle,

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 упо

1.10

 зачем

1.08

vie

1.06

 fume

1.03

av

1.03

 начи

1.03

 Chile

1.02

mee

1.01

 энциклопедия

1.00

 emoc

1.00

POSITIVE LOGITS

 concatenated

1.34

<0xB3>

1.29

 および

1.28

 coupled

1.26

 concatenation

1.25

 sizable

1.22

ackerel

1.21

এবং

1.20

そして

1.20

इयों

1.18

Activations Density 0.039%