INDEX

Explanations

most appropriate answer

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pleasant

1.70

ƾ

1.68

 melancholy

1.62

 excruciating

1.61

த்ரே

1.59

Dalam

1.59

 gloom

1.58

ྛ

1.58

Tah

1.58

 الأوروب

1.57

POSITIVE LOGITS

↵↵↵↵

1.31

↵↵↵↵↵↵↵↵↵↵↵↵↵↵

1.05

↵↵↵↵↵↵↵↵↵↵↵↵

1.03

↵↵↵↵↵↵

1.00

↵↵↵↵↵↵↵↵↵↵

1.00

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.98

↵↵↵↵↵↵↵

0.98

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.97

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.96

Activations Density 3.174%