INDEX

Explanations

proper nouns and concepts

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

believe

-1.22

 honn

-1.22

ᕱ

-1.19

 stroj

-1.17

 sred

-1.17

 Behör

-1.16

 berangkat

-1.16

 srdce

-1.16

-1.15

kään

-1.14

POSITIVE LOGITS

is

2.84

 will

2.30

was

2.02

are

2.02

 should

1.97

 shouldn

1.75

has

1.74

 still

1.66

 must

1.64

 would

1.63

Activations Density 0.042%

No Known Activations