INDEX

Explanations

technical components and systems

The neuron specializes in detecting technical hardware component terms—i.e., the names of mechanical or structural parts in patent‐style descriptions.

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

 flesta

-1.07

 cucchiai

-1.03

 okol

-0.98

idk

-0.98

 kated

-0.96

 drawback

-0.95

 ähnlichen

-0.95

ДО

-0.94

Sole

-0.93

 Füßen

-0.90

POSITIVE LOGITS

째

0.99

 while

0.93

 тре

0.91

途径

0.91

 Möglichkeit

0.88

 части

0.87

 tandis

0.85

 quando

0.85

 जबकि

0.85

˾

0.84

Activations Density 0.123%