INDEX

Explanations

number sequences

The neuron flags multi-digit numeric tokens (especially three or more digits), such as large numbers in citations or section headings.

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

of

-2.45

and

-2.09

-2.08

.’

-1.84

↵↵

-1.81

’,

-1.77

pielen

-1.73

 один

-1.71

to

-1.70

We

-1.70

POSITIVE LOGITS

镡

2.20

 shove

2.05

izes

1.98

fördert

1.98

ες

1.95

vantaged

1.91

 Then

1.90

at

1.88

èrement

1.87

 maravilh

1.84

Activations Density 0.017%