INDEX

Explanations

Conditional logic constructs

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

 where

-1.48

সূত্র

-1.30

any

-1.26

 which

-1.24

 farol

-1.23

 using

-1.18

 Konkur

-1.18

Có

-1.17

与此同时

-1.16

 ہیں

-1.15

POSITIVE LOGITS

It

1.52

However

1.40

𝐾

1.35

 zamiast

1.34

过程

1.33

The

1.32

 вместо

1.27

According

1.26

Instead

1.22

⡯

1.22

Activations Density 0.001%