INDEX

Explanations

acquiring information/actions

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 taking

-1.20

 Taking

-1.13

Taking

-1.08

 takes

-0.87

taking

-0.82

 take

-0.80

er

-0.76

 tomando

-0.71

 takers

-0.68

tak

-0.66

POSITIVE LOGITS

 pleaſure

0.72

 poffe

0.68

TemporalType

0.66

ffions

0.66

 leaſt

0.65

 ſmall

0.64

eace

0.63

.*")]

0.63

__).

0.63

 poffible

0.63

Activations Density 0.111%