INDEX

Explanations

division into three

The neuron detects numeric expressions—digits or spelled‐out numbers indicating counts or quantities.

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

อีก

-1.27

-1.22

də

-1.20

 FIFTH

-1.20

 otro

-1.19

性

-1.19

任何

-1.18

都会

-1.17

 भी

-1.16

 también

-1.15

POSITIVE LOGITS

 tři

1.63

 трех

1.57

 трёх

1.47

 that

1.41

jesz

1.38

 trois

1.38

 trzech

1.33

᷆

1.30

 üç

1.30

 three

1.29

Activations Density 0.248%