INDEX

Explanations

people, groups, or service recipients

This neuron detects numeric data or quantitative expressions (counts, measurements, statistics) in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 because

-1.91

 they

-1.88

because

-1.69

for

-1.52

ｅ

-1.48

 although

-1.38

 until

-1.33

↵

-1.33

 Because

-1.30

 bandeau

-1.29

POSITIVE LOGITS

鼱

1.71

 capucha

1.65

 zwłaszcza

1.63

ͫ

1.61

enablog

1.60

去吧

1.59

 sudadera

1.58

1.56

gives

1.54

colgante

1.52

Activations Density 0.058%