INDEX

Explanations

time units and durations

The neuron flags numerical expressions—including digits or spelled‐out numbers and their associated units (e.g. years, percent, hours).

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

 उच्च

-0.89

嘉

-0.82

дний

-0.82

 smokey

-0.82

 narcissistic

-0.81

：

-0.80

fileupload

-0.80

 бъ

-0.80

定価

-0.79

 Derfor

-0.79

POSITIVE LOGITS

 thuy

1.17

 EAST

1.07

 BETTER

1.05

 WONDER

1.02

east

0.93

 behov

0.92

اخته

0.91

 serpiente

0.90

 نعم

0.90

kowa

0.90

Activations Density 0.094%