INDEX

Explanations

The neuron strongly activates on code tokens defining or referencing hardware registers, control bits, and related macro constants.

register names and code bits

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

когда

-1.25

getTotal

-1.02

mentos

-1.02

 velkommen

-0.94

డా

-0.94

 invade

-0.93

ники

-0.91

 Griechenland

-0.89

 pueden

-0.89

//

-0.88

POSITIVE LOGITS

しなかった

1.21

が悪い

1.13

 Lopes

1.04

焢

1.00

lanti

1.00

 remarks

1.00

ÉM

0.98

が上が

0.98

andExpect

0.98

 Octobre

0.98

Activations Density 0.006%