INDEX

Explanations

studies showing results

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

 then

-1.68

 different

-1.38

人也

-1.33

 Also

-1.27

not

-1.23

 also

-1.22

 needed

-1.20

 alongside

-1.19

Hvad

-1.18

 Alongside

-1.17

POSITIVE LOGITS

 which

3.48

who

2.69

 που

2.39

 که

2.34

 where

2.25

ที่

2.22

ซึ่ง

2.03

 wherein

2.02

 الذي

2.00

 которая

1.95

Activations Density 0.063%