INDEX

Explanations

Those

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

RegressionTest

-0.89

PerformLayout

-0.69

MessageOf

-0.68

AddTagHelper

-0.66

Filmografie

-0.66

 للمعارف

-0.65

 للاسماء

-0.64

ыгана

-0.63

</tfoot>

-0.60

رشف

-0.59

POSITIVE LOGITS

 remember

0.57

 would

0.50

 unaware

0.48

gunta

0.47

 refer

0.46

 wondering

0.45

 remembers

0.43

 wonder

0.42

 rebuke

0.41

 henvisninger

0.41

Activations Density 0.002%