INDEX

Explanations

offers and solicitations

The neuron flags lines that look like classified‐ad or help‐wanted headers—short, all‐caps job or recruitment notices.

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

↵↵↵↵↵↵↵↵↵↵↵

-1.33

↵↵↵↵↵↵↵↵

-1.32

 ........

-1.29

iamo

-1.26

).—

-1.24

̷

-1.23

 .......

-1.20

↵↵↵↵↵↵↵

-1.19

neous

-1.18

-1.13

POSITIVE LOGITS

for

1.49

 ингреди

1.26

！

1.22

at

1.21

in

1.21

 дуже

1.20

Those

1.18

🩱

1.17

 extraordinarily

1.17

six

1.15

Activations Density 0.005%