INDEX

Explanations

`#define` or `\newcommand`

The neuron fires on formal legal/technical definition keywords (e.g. “defined,” “herein,” “referred,” “abbreviated,” “used”) that typically appear in specification or statute definition sections.

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

 أنه

0.89

没有人

0.82

 начали

0.79

 إنه

0.78

 производи

0.77

 отсутствует

0.77

仕様

0.75

 Doppio

0.75

exécution

0.75

 অর্থাৎ

0.74

POSITIVE LOGITS

 istilah

0.79

una

0.75

 words

0.74

 palabras

0.74

orat

0.72

apo

0.70

 sights

0.70

scaping

0.70

 mart

0.69

 substituents

0.69

Activations Density 0.042%