INDEX

Explanations

research and study language

The neuron is selectively activated by numeric tokens—especially numbers, percentages, table or figure references, and other digit‐based data.

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

 掃除

-2.52

 chré

-2.30

皈

-2.23

 パソコン

-2.11

Eſ

-2.05

 яйцо

-1.98

 eorum

-1.94

 Theologie

-1.91

 religieuses

-1.84

辍

-1.84

POSITIVE LOGITS

ֶּ

2.75

ַּ

2.41

 Selasa

2.39

ְּ

2.23

 Sabtu

2.22

ֵּ

2.09

ִּ

1.98

 مباراة

1.96

简约

1.88

ּוֹ

1.88

Activations Density 8.102%