INDEX

Explanations

particular space, Canadian Flag

likelihood and response prediction

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 сущ

0.50

وری

0.45

ема

0.45

ं

0.45

 восприя

0.45

эль

0.44

μαν

0.44

 asociaciones

0.43

باد

0.43

िको

0.42

POSITIVE LOGITS

az

0.51

gment

0.49

 commande

0.48

ka

0.48

 somehow

0.47

 handmade

0.47

()'

0.47

ed

0.46

we

0.46

ROC

0.46

Activations Density 0.002%