INDEX

Explanations

sorry, regret, protocols, enforcement

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

 amazement

1.48

 안전

1.37

겁

1.37

 이슈

1.36

 firef

1.35

 굉장히

1.33

elhos

1.33

 ezingu

1.32

직

1.31

 kudos

1.31

POSITIVE LOGITS

未能

1.25

欠

1.24

無法

1.19

loss

1.15

No

1.10

残念

1.10

 отсутствие

1.09

Loss

1.08

不能

1.05

absence

1.05

Activations Density 0.195%