INDEX

Explanations

from the

The neuron specializes in spotting numeric expressions (especially numbers with decimals or grant/reference figures) embedded in 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

これが

-1.64

癭

-1.59

 lindos

-1.50

躥

-1.43

deinit

-1.43

今回の

-1.42

筴

-1.41

 Commencez

-1.41

嗐

-1.40

 critères

-1.39

POSITIVE LOGITS

of

2.47

by

2.05

the

1.88

but

1.51

 From

1.48

 Чтобы

1.48

not

1.47

 from

1.44

↵

1.40

 other

1.39

Activations Density 0.115%