INDEX

Explanations

on

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ä¸ªçĻ¾åĪĨçĤ¹

-0.26

ieren

-0.25

 categoryId

-0.24

cken

-0.24

 EINA

-0.24

å«©

-0.24

pit

-0.23

ibern

-0.23

lear

-0.23

ForResource

-0.23

POSITIVE LOGITS

VERY

0.30

very

0.30

CCA

0.26

vida

0.26

éĽ¶åĶ®

0.25

 lady

0.25

uty

0.24

 Lady

0.24

qual

0.24

å®¡

0.24

Activations Density 0.002%