INDEX

Explanations

approximation of integral

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

Ð¾ÑģÑĥÐ´

-0.11

Ð¸ÑĨÐ¸Ð½

-0.10

Ð¼ÐµÑĨÑĮ

-0.10

ÐµÑĢÐµÑĩ

-0.10

ÑĦÐµÐº

-0.10

 Serg

-0.10

Ð¾Ð³Ð¾Ñİ

-0.10

luluk

-0.10

Ð¼ÐµÑĢÐ¸Ðº

-0.09

ÑĨÐµÑģ

-0.09

POSITIVE LOGITS

Ñģ

0.14

Ñİ

0.11

Ð±

0.11

Ñĭ

0.10

Ð¾

0.10

0.09

 guar

0.09

Ð¡

0.09

 Ñģ

0.09

Activations Density 0.187%