INDEX

Explanations

required, recommended, commonly

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 technology

-0.11

 technologies

-0.10

 engineered

-0.10

rafted

-0.09

adt

-0.09

	internal

-0.09

ered

-0.09

 mechanisms

-0.08

odega

-0.08

 Platt

-0.08

POSITIVE LOGITS

 recommended

0.12

 reportedly

0.10

 allegedly

0.10

 supposed

0.10

ips

0.10

recommended

0.10

 supposedly

0.10

 recommendation

0.10

 either

0.09

 commonly

0.09

Activations Density 0.081%