INDEX

Explanations

laws, licenses, and places

The neuron fires on numeric tokens—i.e. any standalone digits or version/year numbers.

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

 then

-1.48

-1.20

-1.19

 então

-1.18

-1.15

</strong>

-1.14

</b>

-1.13

</em>

-1.12

Â

-1.12

POSITIVE LOGITS

內容簡介

1.52

 conster

1.45

 confr

1.41

 Sempre

1.34

 ビキニ

1.31

Ilustra

1.30

 notor

1.30

Программа

1.30

Тема

1.27

 distanciation

1.24

Activations Density 0.131%