INDEX

Explanations

"Ð»ÑİÐ±Ð°Ñı Ð·Ð°Ð¼ÐµÐ½Ð°"

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

 Ð¸Ð¼Ñĥ

-0.15

 ÐµÑģÑĤÐµ

-0.14

 Ð¾Ð³ÑĢÐ°

-0.14

 Ð·Ð°Ð²Ð¸ÑģÐ¸Ð¼

-0.14

 Ð±ÑĥÐ¼Ð°

-0.14

 Ð¿ÑĢÐ¾ÑĦÐµÑģÑģÐ¸Ð¾Ð½Ð°Ð»ÑĮ

-0.13

 ÑĤÐ²Ð°

-0.13

 Ð·Ð°ÑıÐ²

-0.13

 ÐµÐ»ÐµÐº

-0.12

 ÐŀÐ´Ð½Ð°

-0.12

POSITIVE LOGITS

 Ð½Ð°

0.11

âĦĸâĦĸ

0.11

 Ð²

0.10

sage

0.09

 ÑĨÐ¸Ð¼

0.09

aina

0.09

owe

0.09

Ð½Ð°Ð´Ð»ÐµÐ¶

0.09

 territory

0.08

 sage

0.08

Activations Density 0.129%