INDEX

Explanations

code syntax or programming initialization

This neuron activates on numeric tokens (especially multi‐digit and decimal numbers, like timestamps or floating-point values).

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

द्रा

0.57

 celebró

0.55

utup

0.54

rouw

0.54

াহি

0.54

 литератур

0.54

zor

0.53

zu

0.52

 قام

0.52

ards

0.51

POSITIVE LOGITS

}\

0.66

 ইনি

0.64

========

0.62

}}\

0.59

 ใจ

0.58

 Immac

0.58

==============

0.57

}=(-

0.55

➖➖

0.54

}]\

0.54

Activations Density 0.004%