INDEX

Explanations

groups and demographics

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

-0.09

wich

-0.09

 squadron

-0.09

 Wend

-0.09

sik

-0.09

 subdivisions

-0.08

 subsidiaries

-0.08

 Ensemble

-0.08

jab

-0.08

POSITIVE LOGITS

 groups

0.46

 group

0.40

groups

0.32

 Groups

0.30

 grupos

0.30

 population

0.30

 segment

0.30

 nhÃ³m

0.29

Groups

0.28

 grupo

0.28

Activations Density 0.176%