INDEX

Explanations

optional or nullable

The neuron fires on tokens indicating optional or nullable types—e.g. “Maybe,” “Nullable,” or other null‐related type markers.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ܜ

-1.61

岨

-1.55

iserie

-1.55

呲

-1.52

ïque

-1.52

磻

-1.52

 semblent

-1.50

 necessárias

-1.48

ulement

-1.45

۠

-1.43

POSITIVE LOGITS

of

2.70

you

1.69

 addition

1.61

甚至

1.59

This

1.55

 further

1.55

 هذه

1.52

 incluso

1.52

1.50

unakan

1.50

Activations Density 0.002%