INDEX

Explanations

capability or consequence

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

Becoming

-1.21

 Becoming

-1.18

becoming

-1.16

 becoming

-1.15

become

-1.09

เป็น

-1.03

 become

-0.98

 Become

-0.97

became

-0.96

becomes

-0.90

POSITIVE LOGITS

 achieve

1.54

 obtain

1.30

 achieved

1.27

 have

1.27

 expect

1.27

 need

1.26

end

1.16

 dealing

1.07

 then

1.05

 achieving

1.05

Activations Density 0.237%