INDEX

Explanations

same virus, evolving virus

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

 Infect

-0.18

 infectious

-0.17

 infected

-0.12

 contagious

-0.12

 inoc

-0.12

 infect

-0.12

 vaccinated

-0.10

 microbial

-0.10

 outbreaks

-0.10

é±

-0.10

POSITIVE LOGITS

 caus

0.15

 coron

0.13

 strain

0.13

 viruses

0.13

 envelop

0.12

 virus

0.12

 causing

0.12

 Enter

0.12

 strains

0.12

onavir

0.12

Activations Density 0.060%