INDEX

Explanations

commenced as, fueled by, what was

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

-schema

-0.12

 activism

-0.12

 revolutions

-0.11

Ð½Ð³

-0.10

 revolution

-0.10

 unde

-0.10

 Revolution

-0.10

 reput

-0.09

 Disqus

-0.09

 rulers

-0.09

POSITIVE LOGITS

 demands

0.20

gal

0.16

 demanding

0.16

 demand

0.15

è¯ī

0.12

demand

0.12

 Demand

0.12

 pent

0.12

 ÄĳÃ²i

0.11

è¨´

0.11

Activations Density 0.064%