INDEX

Explanations

The neuron is triggered by sentence-initial tokens (i.e. words or punctuation at the beginnings of sentences).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

помним

0.89

ள்ளனர்

0.75

 расскажем

0.75

 وقلنا

0.70

 тях

0.69

 нього

0.66

 знаем

0.66

 mamy

0.65

tzv

0.65

给我们

0.63

POSITIVE LOGITS

私は

4.28

 tôi

4.22

 আমি

4.16

我在

3.98

 मैं

3.97

我的

3.86

 मुझे

3.84

 mijn

3.83

 मैंने

3.81

 моей

3.78

Activations Density 6.077%