INDEX

Explanations

code creation dates

The neuron detects numeric tokens in date stamps (e.g. parts of “MM/DD/YY” or “YYYY/MM/DD”) in code comments.

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

in

-2.02

 these

-1.95

if

-1.73

any

-1.50

 этих

-1.42

 ellos

-1.41

 khusus

-1.40

for

-1.40

 should

-1.39

 ilyen

-1.38

POSITIVE LOGITS

grandchildren

1.63

🫂

1.56

ícias

1.52

logías

1.50

CLS

1.48

ності

1.43

 coar

1.40

...",

1.39

 всё

1.38

 hvad

1.38

Activations Density 0.050%