INDEX

Explanations

numeric or symbolic representations in the text

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

 https

-0.08

âĢ¯

-0.08

=https

-0.07

https

-0.07

 COVID

-0.07

 âģ

-0.07

 Covid

-0.07

 coronavirus

-0.07

 ðĿ

-0.07

POSITIVE LOGITS

 Reactive

0.11

 reactive

0.10

 Defensive

0.10

/react

0.10

react

0.10

 react

0.10

-react

0.09

 proactive

0.09

React

0.09

 defensively

0.09

Activations Density 0.003%