INDEX

Explanations

otherwise or exceptions

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

ãĢĢï¾Ĭ

-0.09

Mou

-0.09

æĬľ

-0.09

oke

-0.09

elli

-0.09

quin

-0.09

ommen

-0.08

REE

-0.08

dil

-0.08

iem

-0.08

POSITIVE LOGITS

 otherwise

0.62

otherwise

0.49

 Otherwise

0.48

Otherwise

0.41

 OTHERWISE

0.41

 jinak

0.31

 sonst

0.29

åĲ¦

0.27

 wouldn

0.26

 else

0.21

Activations Density 0.069%

No Known Activations