INDEX

Explanations

copyright

The neuron fires on quantitative details—numbers, counts, measures, and their associated units—in the text.

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

Or

-0.83

 twelve

-0.82

Or

-0.78

or

-0.74

 Twelve

-0.73

Twelve

-0.69

𝓽

-0.68

ゝ

-0.68

 potty

-0.68

或

-0.68

POSITIVE LOGITS

∈

0.77

wrt

0.74

но

0.73

atc

0.70

 Staats

0.68

pace

0.68

zola

0.67

zhong

0.66

 Presiding

0.66

 siyah

0.66

Activations Density 0.531%