INDEX

Explanations

your [feedback/request/gratitude]

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

STRUCTION

-1.85

 then

-1.67

じゃなくて

-1.61

apropri

-1.52

 möglichst

-1.44

珢

-1.41

lative

-1.40

irected

-1.39

eapon

-1.38

NOT

-1.37

POSITIVE LOGITS

</h3>

2.67

↵↵↵

1.88

 considérons

1.46

</strong>

1.43

ต้องการ

1.41

↵↵↵↵

1.35

 mortos

1.35

1.33

and

1.33

，“

1.32

Activations Density 0.008%