INDEX

Explanations

for offenses and wrongdoing

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

ehr

-0.12

 grace

-0.11

(Have

-0.11

PH

-0.10

caret

-0.10

rys

-0.10

 Barker

-0.10

orta

-0.10

unh

-0.10

adan

-0.10

POSITIVE LOGITS

 crimes

0.21

 Crimes

0.17

 mistakes

0.16

 offenses

0.15

 sins

0.15

 misconduct

0.15

 actions

0.15

 violations

0.14

 offences

0.14

 violation

0.14

Activations Density 0.104%