INDEX

Explanations

difficulties and their outcomes

contrast and discourse markers

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

因为

0.48

 ਅਤੇ

0.47

 क्योंकि

0.46

và

0.45

 અને

0.43

 এবং

0.43

 ஏனெனில்

0.43

 karena

0.42

 thấy

0.41

 koska

0.41

POSITIVE LOGITS

 אך

0.48

但也

0.45

でしたが

0.44

했지만

0.44

이지만

0.43

 BETTER

0.42

 reimag

0.42

BUT

0.41

हमने

0.41

 übrigens

0.39

Activations Density 0.006%