INDEX

Explanations

past origins

training and learning

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

 ensuing

0.43

곧

0.40

 soon

0.39

Soon

0.39

不久

0.38

 snart

0.38

糗

0.38

 последу

0.38

ందుకు

0.38

 subsequent

0.38

POSITIVE LOGITS

 formerly

1.25

formerly

1.18

 originally

1.12

 Formerly

1.11

 autrefois

1.08

Formerly

1.05

previously

1.02

originally

1.02

 previously

1.00

 Originally

0.95

Activations Density 0.132%