INDEX

Explanations

geographical locations and multinational contexts

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

 --------------------------------------------------------------------------\n

-0.09

 Svens

-0.09

ayd

-0.09

 FOREIGN

-0.08

ansson

-0.08

 Olsen

-0.08

oster

-0.08

elin

-0.08

 ----------------------------------------------------------------------------\n

-0.08

aska

-0.08

POSITIVE LOGITS

 worldwide

0.28

UK

0.24

USA

0.23

 anywhere

0.22

US

0.21

 countries

0.21

 outside

0.20

 across

0.20

 Ã¼lkemiz

0.19

 Europe

0.18

Activations Density 0.458%