INDEX

Explanations

for the

The neuron detects numeric expressions (especially decimal or floating-point numbers).

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

 that

-1.84

 although

-1.40

 защото

-1.33

 especially

-1.32

 money

-1.32

 ponieważ

-1.31

 matemática

-1.30

 young

-1.25

 kuiten

-1.23

 while

-1.23

POSITIVE LOGITS

 trouvera

1.36

 mesma

1.30

 pères

1.30

蔟

1.26

IORITY

1.24

︲

1.24

ACCIÓN

1.23

}-*/;

1.22

蚰

1.22

autoreleasepool

1.20

Activations Density 0.117%