INDEX

Explanations

legal appeals and petitions

The neuron activates on long runs of whitespace or repeated non‐alphanumeric characters (e.g. extended space blocks or separator lines).

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

and

-1.66

 four

-1.24

 burgeoning

-1.20

for

-1.16

 hombres

-1.16

on

-1.15

 revital

-1.13

 แท้

-1.13

srd

-1.12

 gave

-1.10

POSITIVE LOGITS

 largely

1.59

 !!!!

1.47

 remarkably

1.47

 ゆう

1.34

 !!!!!

1.27

 blatant

1.27

 exceptionally

1.24

 VERY

1.20

 hugely

1.20

 significantly

1.19

Activations Density 0.014%