INDEX

Explanations

courtesy of

The neuron detects image or media credit attributions, especially “Courtesy of …”–style tokens.

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

 Appellants

-1.63

 aktiven

-1.48

 günstiger

-1.47

骉

-1.45

ギュア

-1.45

 divertida

-1.45

 unveil

-1.44

 همه

-1.42

 continuare

-1.42

 使っ

-1.41

POSITIVE LOGITS

蒽

1.91

 from

1.89

者的

1.87

 ikke

1.81

of

1.76

umpulkan

1.74

 only

1.73

 although

1.73

 namun

1.71

 them

1.69

Activations Density 0.003%