INDEX

Explanations

frequent occurrences or assertions

often referred to

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 постоянно

0.75

 constantly

0.73

始终

0.70

常に

0.70

 завжди

0.70

 எப்போதும்

0.67

 always

0.65

 всегда

0.63

 luôn

0.62

 دائما

0.61

POSITIVE LOGITS

 бывает

1.19

 times

1.04

 اوقات

1.03

 overlooked

1.02

times

0.91

 cited

0.89

 misunderstood

0.88

 referred

0.86

 overlook

0.79

referred

0.77

Activations Density 0.183%