INDEX

Explanations

forms of 'to be' followed by descriptor

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

 Pradesh

-1.29

il

-1.29

--;

-1.28

illi

-1.26

ške

-1.24

},

-1.23

垕

-1.21

 forKey

-1.20

or

-1.19

什么事

-1.18

POSITIVE LOGITS

but

1.40

既に

1.27

:”

1.25

1.21

燗

1.16

 jovens

1.11

を選

1.10

すでに

1.09

нятие

1.09

 anún

1.07

Activations Density 0.031%