INDEX

Explanations

its code/object

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

 Ð¾ÑĢÐ³Ð°Ð½Ð¸Ð·Ð°

-0.15

 ÑįÐºÑģÐ¿Ð»ÑĥÐ°ÑĤÐ°

-0.13

 Ð¸Ð½ÑĦÐ¾ÑĢÐ¼Ð°

-0.13

 Ð³ÐµÐ½ÐµÑĢÐ°

-0.13

 Ð¾Ð±Ð»Ð°ÑģÑĤÑĮ

-0.13

 Ð¿Ð»Ð¾ÑīÐ°Ð´

-0.11

 à¤ľà¤Ĺà¤¹

-0.10

 ÑĨÐµÑĢÐº

-0.10

 Ð²Ð¸ÑıÐ²

-0.10

 Ð±Ð¾ÑĢÑĮ

-0.10

POSITIVE LOGITS

chodu

0.13

nika

0.12

Ð°Ð»ÑĮÐ½Ð¾Ð³Ð¾

0.12

boru

0.12

ÐµÐ½Ð½Ð¾Ð³Ð¾

0.12

ovanÃ©ho

0.12

ÎµÎ¯Î¿Ïħ

0.12

Î¯Î¿Ïħ

0.12

Î¼Î±ÏĦÎ¿ÏĤ

0.12

ÑĩÐ½Ð¾Ð³Ð¾

0.12

Activations Density 0.075%