INDEX

Explanations

error handling and messages

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

 appartenant

1.20

 appartiennent

1.13

 integrante

1.08

 provoqu

1.07

 ennen

1.06

 pertence

1.06

acteur

1.05

 pertenc

1.05

 acteur

1.04

 членом

1.04

POSITIVE LOGITS

 options

1.66

 option

1.57

可以选择

1.55

可选

1.49

 Options

1.40

 opción

1.40

 configurations

1.31

 opciones

1.30

options

1.29

 Option

1.29

Activations Density 0.585%