INDEX

Explanations

final answer boxed

The neuron fires on non-Latin script fragments (e.g. Thai characters), i.e. tokens containing foreign-language Unicode characters.

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

"><

0.81

}}}{

0.63

∣

0.62

'<

0.61

|<

0.61

0.59

"<

0.58

"/><

0.58

 đẳng

0.57

!}\

0.57

POSITIVE LOGITS

 care

0.63

 দেখাতে

0.61

 Empower

0.61

 केयर

0.61

 दिखाना

0.60

 сотрудни

0.59

 läh

0.59

cture

0.58

airobi

0.58

ennie

0.58

Activations Density 0.003%