INDEX

Explanations

check the

The neuron strongly activates on instances of “check” (and its variants like “checking,” “double-check,” etc.), i.e. words signaling verification or inspection.

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

 குறிப்பி

0.67

แก่

0.60

 réussite

0.59

 सच्ची

0.59

 erhe

0.57

 "'.$

0.57

ணமாக

0.57

 haupt

0.56

 bezeichnet

0.56

 उत्साहित

0.55

POSITIVE LOGITS

是否有

1.66

是否

1.27

有没有

1.27

是否存在

1.27

有沒有

1.27

 checking

1.21

是不是

1.20

checking

1.19

 thoroughly

1.18

看看

1.17

Activations Density 0.642%