INDEX

Explanations

special characters

The neuron fires strongly on programming‐style symbols and tokens—like operators (e.g. “=”, “|”, “==”, “?”), punctuation, parentheses, and numeric literals—so it’s identifying fragments of source code.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

"";

1.38

[];

1.25

"",

1.18

 "");

1.18

 Integer

1.14

{};

1.12

 null

1.09

"")

1.09

 parseInt

1.09

[]

1.02

POSITIVE LOGITS

झ्या

0.66

approche

0.65

逾

0.64

ेंद्रीय

0.64

तीच्या

0.64

收购

0.63

 स्वातंत्र

0.62

Conservation

0.62

Tá

0.61

évolution

0.61

Activations Density 0.001%