INDEX

Explanations

cultural and social diversity

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

Negative Logits

bif

-0.11

 Latino

-0.09

 rout

-0.09

ighth

-0.09

çĽĳåĲ¬é¡µéĿ¢

-0.09

 Hispanic

-0.08

 gratuiti

-0.08

ControlItem

-0.08

Erd

-0.08

/testify

-0.08

POSITIVE LOGITS

 diversity

0.32

 Diversity

0.24

 melting

0.24

 tolerance

0.22

 diverse

0.21

 divers

0.20

div

0.19

iversity

0.18

 plural

0.17

 tolerant

0.17

Activations Density 0.087%

No Known Activations