INDEX

Explanations

methods, techniques, or processes

This neuron detects mentions of scientific or technical procedures—words like “method,” “technique,” or “type” (often in section headings or document structure).

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

and

-1.77

was

-1.52

]}{

-1.38

as

-1.37

 mohl

-1.34

 while

-1.34

maları

-1.30

ıyordu

-1.30

cipy

-1.29

 before

-1.28

POSITIVE LOGITS

 любой

1.43

you

1.35

 beendet

1.34

 préalable

1.34

 buatan

1.34

 verhindert

1.33

three

1.27

any

1.27

Activations Density 0.222%