INDEX

Explanations

negative emotions, errors, and bad situations described in the text.

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

Portail

-0.89

>",

-0.70

__':

-0.69

 virke

-0.68

 plenamente

-0.66

 EconPapers

-0.66

CPtr

-0.65

 GetEnumerator

-0.64

 насеље

-0.64

adaptiveStyles

-0.63

POSITIVE LOGITS

 damage

0.95

 Worse

0.93

 worse

0.92

 injured

0.92

 badly

0.91

 DAMAGE

0.91

Worse

0.90

 malfunction

0.89

bad

0.88

 injuring

0.87

Activations Density 6.589%