INDEX

Explanations

kept busy, delayed, or distracted

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

of

-1.55

 that

-1.47

 persönliche

-1.36

 Already

-1.33

 Urteil

-1.30

stealth

-1.28

 vorbere

-1.27

selbe

-1.26

 Hinblick

-1.24

 zuneh

-1.23

POSITIVE LOGITS

</strong>

1.72

();

1.64

 innan

1.59

1.55

 questi

1.52

ización

1.51

usions

1.49

varepsilon

1.48

 súper

1.48

Activations Density 0.021%