INDEX

Explanations

subtraction and negative numbers

The neuron activates on numeric tokens and expressions—i.e. digits, numbers, and related numeric values.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

[++

1.02

(++

0.91

 افزایش

0.78

増加

0.78

 resurrected

0.76

 વધ

0.73

 meningkat

0.73

เพิ่ม

0.72

增

0.71

 tăng

0.70

POSITIVE LOGITS

 minus

2.58

 subtract

2.55

 subtracting

2.47

subtract

2.47

减

2.30

 subtraction

2.27

minus

2.25

 subtracted

2.25

Subtract

2.23

 Minus

2.20

Activations Density 0.267%