INDEX

Explanations

predicts continuation

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

Ð½ÑĥÑĤÑĮÑģÑı

-0.11

Tet

-0.11

MT

-0.10

ings

-0.10

fe

-0.09

ÑĩÐ¸ÑģÐ»

-0.09

ÐµÑĤÑĮÑģÑı

-0.09

 cock

-0.09

Ð»Ð°Ð³Ð°

-0.09

 centr

-0.09

POSITIVE LOGITS

Ã¡vÃ¡

0.14

ÅĪuje

0.13

Î»Î±Î¼Î²

0.12

 à¤ķà¤°à¤¤

0.12

ÑĭÐ²Ð°ÐµÑĤ

0.12

Ã¡vajÃŃ

0.12

Ð²Ð¾Ð»Ñı

0.12

ÑĢÐ°Ð±Ð°ÑĤ

0.11

ÏģÎ¯Î¶

0.11

à¤¿à¤²à¤¤

0.11

Activations Density 0.110%