INDEX

Explanations

code and structure

The neuron detects numeric literals (numbers, whether integers or decimals) 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

or

-2.38

as

-2.06

 their

-1.90

at

-1.89

 from

-1.79

one

-1.79

-1.75

by

-1.73

our

-1.69

but

-1.66

POSITIVE LOGITS

 diriger

2.09

 faciliter

1.98

眎

1.83

 teater

1.80



1.78

{}

1.75

 tilla

1.74

..”

1.73

li

1.70

 poë

1.70

Activations Density 0.050%