INDEX

Explanations

That followed by information

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

这些

-1.74

 даже

-1.73

這些

-1.72

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.69

 these

-1.64

↵↵↵↵↵↵↵↵↵↵↵↵

-1.62

</h3>

-1.58

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.56

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.54

 食事

-1.53

POSITIVE LOGITS

 البته

1.60

esomeness

1.44

素晴らしい

1.41

队友

1.37

ꪹ

1.35

回到了

1.33

ing

1.30

非常好

1.30

 ardından

1.30

有些不

1.30

Activations Density 0.018%