INDEX

Explanations

combining roots and suffixes

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 være

1.35

 fordi

1.28

 hvis

1.28

 disse

1.27

 brukes

1.24

 stadig

1.19

 hver

1.19

 flere

1.18

 ikke

1.18

 grunn

1.18

POSITIVE LOGITS

 только

2.78

 его

2.63

 до

2.62

 так

2.58

 не

2.57

 из

2.54

 много

2.52

 очень

2.51

 три

2.48

 ко

2.48

Activations Density 0.021%