INDEX

Explanations

reporting speech or quotes

The neuron fires on the main topical keywords or proper‐noun subjects that signal each article’s focus (e.g. drug names, product titles, organizational or topic nouns).

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

カシ

-1.02

our

-1.02

 coute

-1.02

が見える

-1.02

Our

-1.00

équi

-1.00

 réfrig

-0.98

选择了

-0.96

our

-0.96

 perigo

-0.95

POSITIVE LOGITS

 according

2.03

 said

1.76

 says

1.75

1.62

“

1.55

 volgens

1.36

だそうです

1.32

他說

1.26

"[

1.23

他说

1.20

Activations Density 0.027%