INDEX

Explanations

don't minimize failure

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

友達

1.45

azt

1.45

 różne

1.44

 ďal

1.43

 allons

1.42

 různ

1.41

 filóso

1.39

 തവണ

1.38

 οποίο

1.36

 szere

1.36

POSITIVE LOGITS

度和

1.73

และความ

1.69

पूर्वक

1.62

iousness

1.61

onset

1.46

omeness

1.43

 Sweep

1.41

 पूर्वक

1.41

USB

1.39

ateness

1.38

Activations Density 0.742%