INDEX

Explanations

action-oriented verbiage related to involvement and change

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Negative Logits

ookies

-0.06

enheim

-0.06

 corre

-0.06

kÄ±

-0.06

ÛĮÙĨÙĩ

-0.06

ï¿

-0.06

abr

-0.06

lington

-0.06

attern

-0.06

etas

-0.06

POSITIVE LOGITS

 again

0.09

again

0.08

 elsewhere

0.07

Again

0.07

 last

0.07

0.06

racat

0.06

.ImageAlign

0.06

ritel

0.06

 wherever

0.06

Activations Density 0.047%

No Known Activations