INDEX

Explanations

patent section titles

The neuron fires on numeric tokens—especially decimal numbers and other numeric references (e.g. measurements, patent cites, section figures)—in the technical document.

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

-2.03

 çünkü

-1.96

 všechny

-1.85

 everybody

-1.84

และ

-1.70

ерез

-1.59

 ļoti

-1.58

🫓

-1.58

شته

-1.55

のおかげ

-1.52

POSITIVE LOGITS

 awalnya

1.84

 hidupnya

1.81

趼

1.75

1.70

靱

1.66

 termurah

1.64

 beserta

1.61

 جالب

1.60

ЛИ

1.60

 pelatihan

1.59

Activations Density 0.006%