INDEX

Explanations

variable assignment with numbers

The neuron strongly activates on numeric literals used as indices or arguments in code—e.g. array-access brackets like “[2]” or function-call parentheses like “group(1)”.

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

 diyor

-1.00

 cilvē

-0.90

cambrian

-0.86

outFile

-0.85

IFYING

-0.83



-0.82

寐

-0.82

 tył

-0.82

 añadió

-0.81

などが

-0.81

POSITIVE LOGITS

ayashi

0.91

0.86

phonie

0.84

 arşivlendi

0.82

ią

0.82

 השני

0.82

nloa

0.82

 Gosling

0.81

砾

0.79

 namesake

0.79

Activations Density 0.006%