INDEX

Explanations

Boost Software License

The neuron fires on tokens that mark high‐level code or file declarations—things like “package,” “struct,” project/name identifiers, class names, version numbers, etc.

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

Ἅ

-0.74

衛隊

-0.71

getPost

-0.71

 суть

-0.69

コミュニ

-0.66

myspace

-0.66

Vå

-0.66

ને

-0.66

garde

-0.65

("."

-0.65

POSITIVE LOGITS

putra

0.93

interface

0.88

 interfaces

0.84

Rourke

0.84

athed

0.84

 curse

0.84

 shutil

0.83

grund

0.81

packets

0.80

roke

0.79

Activations Density 0.001%