INDEX

Explanations

language related to break-ins or forced entries

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

Ire

-0.07

obble

-0.07

olleyError

-0.07

istrovstvÃŃ

-0.07

 drib

-0.07

ayment

-0.07

 uphill

-0.07

uet

-0.07

amburger

-0.06

Ä±sÄ±t

-0.06

POSITIVE LOGITS

 access

0.06

 panel

0.06

 window

0.06

 bypass

0.06

 latch

0.06

 Shore

0.06

Cle

0.06

 Access

0.06

 Suff

0.06

 vegetation

0.06

Activations Density 0.010%