INDEX

Explanations

com/ in URLs

The neuron detects top‐level domain tokens (like “com,” “org,” “io,” “net”) in code import or module paths.

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

for

-2.19



-1.58

しているので

-1.57

 prêtres

-1.56

しましたが

-1.55

 collaborated

-1.54

 justiça

-1.53

埏

-1.53

してくれた

-1.52

できるので

-1.52

POSITIVE LOGITS

</h3>

1.93

↵↵

1.91

1.82

ной

1.68

↵

1.55

⚡️

1.49

Cartney

1.45

";

1.45

“

1.43

you

1.38

Activations Density 0.009%