INDEX

Explanations

explaining code improvements

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

 setzte

0.83

建て

0.74

 अभियंता

0.71

 संके

0.71

 Besitzer

0.68

 történ

0.66

 granite

0.65

 Soder

0.62

境

0.62

rimidine

0.61

POSITIVE LOGITS

↵↵↵↵↵↵↵↵↵

1.41

↵↵↵↵↵↵

1.29

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

1.26

↵↵↵↵↵↵↵↵

1.21

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

1.19

↵↵↵↵↵↵↵

1.18

↵↵↵↵↵

1.17

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

1.16

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

1.14

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

1.09

Activations Density 0.026%