INDEX

Explanations

code and punctuation separators

The neuron fires on numeric literals used as repetition bounds in regex quantifiers (e.g. the “m” and “n” inside `{m,n}`).

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

it

-2.52

-2.05

in

-1.97

-1.84

as

-1.83

if

-1.73

to

-1.72

-1.67

 itself

-1.56

POSITIVE LOGITS

 klarer

2.00

 tabak

1.95

шают

1.85

苼

1.84

 этого

1.81

 kunde

1.79

 corriger

1.78

 eşofman

1.78

 fiyat

1.73

genodigd

1.73

Activations Density 0.004%