references to specific numbers, particularly large population counts

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 itſelf

-1.27

 raiſ

-1.10

 houſe

-1.05

 purpoſe

-1.04

 reaſon

-1.03

ſelf

-1.02

 ſche

-1.01

$_"

-1.01

 ―――――

-1.00

 ſeveral

-1.00

POSITIVE LOGITS

0.52

0.50

0.49

and

0.48

0.47

Two

0.47

TO

0.47

Activations Density 0.027%