INDEX

Explanations

help to improve

improving or making better

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

做法

0.41

ﻉ

0.41

 houden

0.40

 ছুই

0.39

 hust

0.39

 मानने

0.38

手数

0.38

承認

0.38

惹

0.38

道歉

0.38

POSITIVE LOGITS

 better

1.48

更好地

1.38

better

1.29

更好的

1.18

 лучше

1.17

 lepiej

1.16

 meglio

1.15

 mieux

1.14

 beter

1.13

 besser

1.13

Activations Density 0.302%