INDEX

Explanations

focus on innovation

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

anter

-0.09

ï¾Ĩï¾Ĩ

-0.09

 Sour

-0.09

åĨµ

-0.08

 Specialty

-0.08

è®¾è®¡åĻ¨

-0.08

inka

-0.08

rett

-0.08

lei

-0.08

POSITIVE LOGITS

 focus

0.22

focus

0.16

 emphasis

0.15

emphasis

0.14

 commitment

0.14

.focus

0.13

phasis

0.12

 combination

0.12

Focus

0.12

 Focus

0.12

Activations Density 0.031%