INDEX

Explanations

statement and discussion reporting

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

who

-1.43

⚣

-1.38

 agresión

-1.30

頷

-1.27

tà

-1.20

 gelt

-1.20

see

-1.19

dù

-1.18

змом

-1.17

幛

-1.17

POSITIVE LOGITS

to

1.27

してしまう

1.21

 brings

1.18

 دوباره

1.18

ₗ

1.17

欧洲

1.16

之后

1.16

。。。。

1.14

 просто

1.13

 گیاه

1.13

Activations Density 0.026%