INDEX

Explanations

negative implications related to accountability and accountability failures in various contexts

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ference

-0.07

 Privacy

-0.06

ump

-0.06

crire

-0.05

nings

-0.05

elligence

-0.05

retty

-0.05

 Regents

-0.05

 regression

-0.05

enas

-0.05

POSITIVE LOGITS

.RightToLeft

0.07

/REC

0.07

clas

0.06

ori

0.06

Tod

0.06

iou

0.06

åŁŁ

0.06

ede

0.06

.sax

0.06

oni

0.06

Activations Density 0.011%