INDEX

Explanations

structural components and frameworks

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

or

-1.49

Maybe

-1.41

to

-1.38

「

-1.29

Because

-1.29

 want

-1.19

でも

-1.17

vál

-1.16

 благодаря

-1.16

exit

-1.16

POSITIVE LOGITS

</code>

1.61

ーめん

1.55

asanjo

1.38

慵

1.33

</h5>

1.33

橚

1.30

ANDREW

1.28

}}$

1.26

 sieben

1.23

一顿

1.23

Activations Density 0.067%