INDEX

Explanations

please and thanks

polite requests and trust

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

激情

0.41

 onlar

0.38

 ebben

0.38

vij

0.38

 laget

0.38

 був

0.37

tarea

0.37

Stadt

0.37

ູນ

0.36

 വിജയ

0.36

POSITIVE LOGITS

 Please

0.49

 please

0.48

 ব্যবস্থ

0.47

 trust

0.46

 deductible

0.46

材质

0.44

 Trust

0.44

请

0.43

let

0.40

 প্রস্তাবে

0.40

Activations Density 0.004%