INDEX

Explanations

C-Reactive Protein levels

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

 Rated

-0.10

-0.09

asant

-0.09

 Rescue

-0.09

 bell

-0.09

 Nurs

-0.08

ffer

-0.08

 Pond

-0.08

kest

-0.08

Res

-0.08

POSITIVE LOGITS

 levels

0.19

ng

0.16

 Creat

0.15

 blood

0.15

Creat

0.15

levels

0.15

 values

0.15

 Levels

0.14

 creat

0.14

Levels

0.14

Activations Density 0.034%