INDEX

Explanations

fraction or proportion of

The neuron detects numerical quantitative expressions—especially proportions, fractions, densities, rates, and other percentage-style measurements—in the text.

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

-1.83

↵↵

-1.72

”-

-1.49

 adquisición

-1.49

and

-1.40

 alojamientos

-1.36

 zajíma

-1.35

 viviendas

-1.34

 competición

-1.34

也知道

-1.33

POSITIVE LOGITS

of

2.41

Danke

1.41

Źródło

1.41

 ISLANDS

1.38

Andere

1.37

Listo

1.35

ährlich

1.34

APPENDICES

1.29

 sanoo

1.26

Secondo

1.25

Activations Density 0.066%