INDEX

Explanations

in-house, in-depth words

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

ോളിക

2.30

 accrue

2.25

 lunghezza

2.23

AppDesigner

2.23

 ভাবেই

2.20

्यर्थ

2.17

িও

2.13

芜

2.06

وَ

2.06

𝚕

2.06

POSITIVE LOGITS

uited

1.81

rolled

1.71

ating

1.68

ning

1.62

ched

1.60

king

1.59

itting

1.58

uit

1.58

aba

1.55

bad

1.55

Activations Density 0.224%