INDEX

Explanations

possibility and necessity

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

-2.95

 they

-2.91

 土産

-2.31

 unmatched

-2.30

和

-2.27

要

-2.25

葙

-2.22

ri

-2.20

</strong>

-2.17

and

-2.16

POSITIVE LOGITS

2.63

 který

2.33

 což

2.25

 было

2.19

 будет

2.16

 kteří

2.08

It

2.05

痦

2.05

 muž

1.98

 dům

1.97

Activations Density 0.003%