INDEX

Explanations

added agreement know suspected revealed

understanding user intent

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 considerando

0.24

 используя

0.24

 ayant

0.23

 щоб

0.22

 Them

0.22

 думаю

0.22

 implying

0.21

 Mostly

0.21

தற்காக

0.21

 teniendo

0.21

POSITIVE LOGITS

<unused2204>

0.28

 they

0.27

 excitedly

0.26

via

0.26

 τελευτα

0.25

abouts

0.25

<unused2169>

0.25

这是一

0.25

 thats

0.24

這是

0.24

Activations Density 0.095%