INDEX

Explanations

performance rather thanreason forworks asgenerated

outcomes and conditions

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 drooping

0.25

 jaoks

0.24

ើរ

0.24

ത്തിയാ

0.24

 হও

0.24

 xaxis

0.24

 dropdown

0.23

㏄

0.23

 tuve

0.23

胿

0.23

POSITIVE LOGITS

 вполне

0.29

从而

0.28

 Reiter

0.28

 thereby

0.27

affirming

0.25

 гораздо

0.24

inspiring

0.24

 Pérez

0.23

 Conversely

0.23

 Vaughan

0.23

Activations Density 0.238%