INDEX

Explanations

adjusted accordingly

The neuron activates on numeric expressions—quantities, percentages, and other statistical figures.

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

有一种

0.64

这是一个

0.62

because

0.61

Todo

0.58

এটা

0.57

ሁሉም

0.57

小さな

0.56

 모든

0.56

ብዙ

0.56

이나

0.55

POSITIVE LOGITS

 respectively

5.56

respectively

4.97

 respectivamente

4.88

 respectivement

4.59

 соответственно

4.56

 respective

4.31

 respectiv

4.22

それぞれ

3.80

respective

3.80

分别

3.59

Activations Density 0.274%