INDEX

Explanations

URLs and web addresses

The neuron detects tokens that are parts of web URLs, especially domain names and related URL segments.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

to

-2.20

 السياس

-1.96

鍼

-1.69

 BSEB

-1.68

 мали

-1.63

 rámci

-1.59

 المسلمين

-1.55

 السلطة

-1.52

 الوزراء

-1.52

 September

-1.51

POSITIVE LOGITS

夘

1.81

 udang

1.78

 thoſe

1.77

 theſe

1.72

 that

1.71

 ngan

1.66

ходил

1.63

 bluz

1.59

 kela

1.56

 katun

1.55

Activations Density 0.180%