INDEX

Explanations

referring to someone and their context

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

 eram

2.03

都很

1.93

 nacionales

1.87

 foram

1.85

 đều

1.83

 оказались

1.82

 nacionais

1.81

 permitem

1.80

 furono

1.79

 erano

1.77

POSITIVE LOGITS

 chooses

1.67

 نفسه

1.46

 ತನ್ನ

1.45

 learns

1.44

 himself

1.39

 attains

1.38

 obtains

1.29

 தனது

1.29

 owns

1.28

 தன்னை

1.27

Activations Density 0.459%