INDEX

Explanations

efficiency, readability, code

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

 aforesaid

2.05

ueless

2.02

rezent

2.00

obot

1.97

 manipulative

1.93

 notoriety

1.93

嘚

1.92

 routed

1.91

 bract

1.91

áže

1.90

POSITIVE LOGITS

我们

0.89

we

0.83

 យើង

0.81

我们可以

0.80

ينا

0.78

我们要

0.76

 resolv

0.74

เรา

0.73

我们的

0.73

 kita

0.72

Activations Density 0.012%