INDEX

Explanations

escolha

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

 nuest

-0.15

hog

-0.11

-0.10

Kai

-0.10

ificaciÃ³n

-0.10

ificaciones

-0.10

die

-0.10

 Nail

-0.09

 hurt

-0.09

MI

-0.09

POSITIVE LOGITS

ifique

0.21

ÑĥÐ¹ÑĤÐµ

0.17

Ð²ÐµÐ´Ð¸ÑĤÐµ

0.15

 tenga

0.14

ujte

0.14

Ð¹ÑĤÐµ

0.13

nÄĽte

0.13

 hÃ£y

0.13

enga

0.13

Ð°Ð¹ÑĤÐµ

0.12

Activations Density 0.119%