INDEX

Explanations

end of

This neuron flags words indicating a document position—especially “end” or “bottom” in phrases like “at the end” or “at the bottom.”

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

AR

0.57

内の

0.54

阂

0.54

内で

0.53

શ

0.52

 içine

0.52

后面的

0.52

dataGridView

0.52

คาเดมี

0.52

 لاءِ

0.52

POSITIVE LOGITS

 thereof

0.96

 ofthe

0.91

end

0.88

ของการ

0.86

 hereof

0.83

OfThe

0.80

ofthe

0.78

 দিকের

0.73

陲

0.73

of

0.71

Activations Density 0.297%