INDEX

Explanations

negative prefixes indicating separation or disagreement

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

 Orta

-0.07

saida

-0.07

Ð¢Ðŀ

-0.07

uracy

-0.07

assi

-0.07

iterals

-0.07

FB

-0.06

uil

-0.06

ullet

-0.06

icher

-0.06

POSITIVE LOGITS

/dis

0.10

dis

0.10

(dis

0.10

.dis

0.10

Dis

0.10

Dis

0.09

-dis

0.09

DIS

0.09

.Dis

0.08

dis

0.08

Activations Density 0.038%