INDEX

Explanations

aka, also known as

The neuron fires on capitalized proper‐name tokens—brand names, tool names, acronyms, organization/place names, and other named‐entity words.

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

 แถม

0.67

anilide

0.62

 있었습니다

0.61

)?;

0.59

ylül

0.59

 révèle

0.59

ynitrite

0.59

 memastikan

0.59

 आकर्षक

0.58

 ideje

0.58

POSITIVE LOGITS

aka

1.35

AKA

1.34

 nicknamed

1.30

 nickname

1.28

 colloqu

1.28

简称

1.22

AKA

1.20

叫做

1.16

called

1.15

 nicknames

1.15

Activations Density 1.017%