INDEX

Explanations

can contribute

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

á»Ĩ

-0.09

 Morr

-0.08

 accom

-0.08

kova

-0.08

 indeb

-0.08

inned

-0.08

udiant

-0.08

achuset

-0.08

swana

-0.08

 Attendance

-0.08

POSITIVE LOGITS

 contribute

0.44

 contrib

0.44

 contributed

0.40

 contribution

0.40

 contributes

0.39

 gÃ³p

0.39

 contributing

0.38

contrib

0.38

Contrib

0.35

 Contrib

0.34

Activations Density 0.161%