INDEX

Explanations

hearing about or telling of

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

ugs

-0.09

inez

-0.09

 unthinkable

-0.09

 outlook

-0.09

ickt

-0.08

.partial

-0.08

ennes

-0.08

ifi

-0.08

fut

-0.08

CreatedAt

-0.08

POSITIVE LOGITS

 existence

0.20

 exploits

0.17

 Exist

0.17

existence

0.15

 exist

0.14

Exist

0.14

 happen

0.13

 exists

0.13

åŃĺåľ¨

0.13

exist

0.13

Activations Density 0.075%