INDEX

Explanations

place name prefixes or beginnings

The neuron strongly activates on tokens from “first attested” and etymology/history lines—particularly year digits and root‐word fragments in their original languages.

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

 cherchés

-1.31

ּוֹ

-1.27

 embreagem

-1.14

 opção

-1.07

ִּ

-1.06

illite

-1.04

かの

-1.04

 siè

-1.02

を含む

-1.01

を持つ

-1.00

POSITIVE LOGITS

or

1.41

或

1.33

}")]

1.33

byl

1.20

 Pokud

1.19

+")

1.19

祂

1.16

邉

1.13

BindView

1.13

 espirituales

1.12

Activations Density 0.022%