INDEX

Explanations

contrast 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

No Comments

Negative Logits

 barbaric

1.34

 reprises

1.20

shakes

1.16

 осуществления

1.16

決して

1.15

 anymore

1.13

 технологий

1.13

 alcun

1.13

 nacionais

1.12

 alcuna

1.11

POSITIVE LOGITS

个

1.42

 elevating

1.41

ראש

1.40

وز

1.36

 گیر

1.35

 자기

1.33

 giỏi

1.29

表

1.25

 addTodo

1.24

 مجموع

1.24

Activations Density 0.000%