INDEX

Explanations

Grand Slam titles

New Auto-Interp

Top Features by Cosine Similarity

Configuration

Prompts (Dashboard)

10,000 prompts, 128 tokens each

Dataset (Dashboard)

lmsys/lmsys-chat-1m

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iren

-0.10

orate

-0.10

 finalists

-0.10

 Prize

-0.09

Pra

-0.09

 unbe

-0.09

 Cream

-0.09

 WARRANT

-0.09

 finalist

-0.09

etics

-0.09

POSITIVE LOGITS

 titles

0.40

 title

0.36

title

0.27

 Titles

0.26

titles

0.25

 tÃŃtulo

0.24

.title

0.23

 Title

0.23

 titre

0.22

_titles

0.22

Activations Density 0.045%