INDEX

Explanations

in order to, so that

The neuron fires on Russian question-framing function words (e.g. “нужно,” “чтобы,” “как”) that signal necessity or ask how to do something.

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

失败

-0.88

สำหรับ

-0.85

ків

-0.84

Ķ

-0.82

 kiệm

-0.82

请选择

-0.81

גרפיה

-0.80

ain

-0.80

 chimes

-0.79

ฬ

-0.79

POSITIVE LOGITS

 чтобы

4.97

 щоб

3.53

to

2.89

aby

2.83

 כדי

2.70

 để

2.67

 żeby

2.50

Чтобы

2.30

 تا

2.28

 чтоб

2.19

Activations Density 0.014%