INDEX

Explanations

list separators and punctuation

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

↵↵

-3.52

↵↵↵

-2.14

↵

-1.92

↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.85

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.77

);

-1.74

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.70

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.68

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.67

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

-1.66

POSITIVE LOGITS

2.31

In

2.16

For

1.93

==

1.91

 This

1.90

At

1.79

As

1.79

So

1.78

On

1.77

1.76

Activations Density 0.130%

No Known Activations