INDEX

Explanations

Java context, file, and imports

This neuron detects numerical constants (integer and floating‐point literals) in the code.

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

]}$

-1.18

 batas

-1.08

 gött

-1.07

の方も

-1.06

}();

-1.05

 aDecoder

-1.00

our

-0.98

ſts

-0.97

heureusement

-0.96

 می‌گو

-0.95

POSITIVE LOGITS

それは

1.08

 защото

0.99

したのは

0.97

itorio

0.95

wide

0.94

 gevaar

0.92

 видел

0.91

pshire

0.90

ったのは

0.88

成的

0.88

Activations Density 0.009%