INDEX

Explanations

discovered bythis gene

The neuron fires primarily on standalone numeric tokens (e.g. years, dimensions, indices) and other non‐word symbols (like apostrophes and punctuation).

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

ächtige

-0.71

ệt

-0.71

 includes

-0.70

 decisión

-0.66

 contour

-0.65

 Reno

-0.65

 хими

-0.65

 Frust

-0.64

vô

-0.64

太后

-0.64

POSITIVE LOGITS

ombia

0.73

牙齿

0.72

latlng

0.71

 совмест

0.69

 forcible

0.68

 설명

0.68

に来

0.67

 綺麗

0.67

Wells

0.66

𝒗

0.66

Activations Density 0.097%