INDEX

Explanations

video source, electronically from

The neuron primarily detects numeric sequences—strings of digits (often multi‐digit technical specifications or measurements).

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

民众

-1.62

 السياس

-1.58

 otras

-1.47

 proy

-1.40

 otros

-1.38

 decorar

-1.37

 food

-1.34

₂+

-1.29

 hatten

-1.28

it

-1.27

POSITIVE LOGITS

madeus

1.79

 deng

1.77

 mahu

1.76

 haji

1.71

 微信

1.67

 游戏

1.66

 laga

1.62

kuh

1.61

 hina

1.60

咂

1.59

Activations Density 0.057%