INDEX

Explanations

Ð¸ Ð¾ÑĤÐ½Ð¾ÑģÑıÑĤÑģÑı

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

 Ð¸Ð¼Ñĥ

-0.22

 Ð½ÐµÐ´ÐµÐ»

-0.21

 ÐµÐ»ÐµÐº

-0.15

 Ð´Ð¾ÐºÑĥÐ¼

-0.15

 Ð¾Ð³ÑĢÐ°

-0.14

 ÐµÐ»ÐµÐ¼

-0.14

 Ð¸Ð½ÑĦÐ¾ÑĢÐ¼Ð°

-0.13

 Ð¿Ð¾Ð±Ð°

-0.13

 Ð´Ð¸Ð²Ð¸

-0.13

 ÑĤÐ²Ð°

-0.12

POSITIVE LOGITS

 representatives

0.09

 Dahl

0.09

vem

0.09

èĨ

0.08

'gc

0.08

flo

0.08

 Maduro

0.08

 ÑĥÐ¼Ð¾Ð²

0.08

 Hakk

0.08

ÙĦÛĮÙĦ

0.08

Activations Density 0.087%

No Known Activations