INDEX

Explanations

numbers and quantities

The neuron is effectively inactive—it does not respond to any tokens and so does not detect any pattern.

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

-2.30

-1.73

↵

-1.54

 terá

-1.44

дают

-1.32

);

-1.28

有事

-1.26

 They

-1.26

’.

-1.24

-1.23

POSITIVE LOGITS

of

2.05

 этому

1.32

遅い

1.30

顔を

1.27

 делу

1.26

ním

1.23

 into

1.21

 poissons

1.21

 Resmi

1.20

 القرار

1.20

Activations Density 0.047%