INDEX

Explanations

predicting trends and outcomes

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

lhe

1.60

 pourront

1.55

каў

1.38

 maior

1.37

とう

1.36

 correctamente

1.36

となり

1.34

 необходимо

1.34

で使用

1.33

 Employ

1.33

POSITIVE LOGITS

 sorta

1.91

boom

1.72

sort

1.69

basin

1.62

ordinating

1.62

crazy

1.62

 shenanigans

1.57

 backdoor

1.57

monster

1.55

disciplinary

1.55

Activations Density 0.330%