INDEX

Explanations

foreign words or phrases

The neuron is strongly triggered by named entities and formal titles—particularly capitalized names, acronyms, multi-word organization titles, and numeric labels (years, series numbers, etc.).

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

one

-1.40

 provide

-1.31

can

-1.27

 providing

-1.24

 while

-1.23

or

-1.23

 making

-1.20

 first

-1.20

use

-1.19

 with

-1.18

POSITIVE LOGITS

 모든

1.73

 fidé

1.61

 allerlei

1.61

 siè

1.55

 assister

1.55

 מכל

1.54

 célè

1.52

 dégust

1.51

 mauva

1.46

 incrí

1.46

Activations Density 0.103%