INDEX

Explanations

future prediction or condition

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

なー

-1.27

 recibido

-1.27

 anadolu

-1.23

ことはない

-1.20

 anuncia

-1.16

 kimdir

-1.14

 отвечает

-1.13

 lately

-1.13

 lasted

-1.10

なあ

-1.10

POSITIVE LOGITS

𖤍

1.22

 vermutlich

1.19

 will

1.14

 Maybe

1.11

 likely

1.08

ourites

1.04

 Probably

1.03

螓

1.02

 Unless

1.02

Blockly

1.01

Activations Density 0.086%