INDEX

Explanations

references to specific names or titles related to performance

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

er

-0.08

ru

-0.07

rig

-0.07

oter

-0.07

plet

-0.07

run

-0.07

Ð¸

-0.07

erce

-0.07

ration

-0.07

 RectTransform

-0.07

POSITIVE LOGITS

anooga

0.09

bom

0.07

itudes

0.07

erson

0.07

ught

0.07

wg

0.07

elier

0.07

adoo

0.07

apult

0.06

Kee

0.06

Activations Density 0.012%