INDEX

Explanations

closing parentheses and commas in code

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

 για

-1.23

 penghar

-1.19

 because

-1.16

You

-1.16

Because

-1.15

because

-1.11

osexuality

-1.11

 through

-1.11

One

-1.10

Still

-1.10

POSITIVE LOGITS

REASONS

1.13

 하고

1.03

INGREDIENT

1.02

出てくる

1.02

 見える

0.98

 ursprünglich

0.98

 NTIS

0.96

쫓

0.96

に入った

0.95

出てきた

0.94

Activations Density 0.004%