INDEX

Explanations

para ligar

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

odon

-0.09

.AddListener

-0.08

 Ð·Ð°Ð²Ð¸ÑģÐ¸Ð¼

-0.08

æĺ¯ä¸ª

-0.08

remen

-0.08

Anc

-0.07

Ø³ÙĪØ¨

-0.07

ä¸ĬãģĮ

-0.07

 ráº±ng

-0.07

POSITIVE LOGITS

 Äĳá»ĥ

0.46

 ÑĩÑĤÐ¾Ð±Ñĭ

0.43

 ÑīÐ¾Ð±

0.33

aby

0.32

ä¸ºäºĨ

0.31

 inorder

0.28

 nháº±m

0.26

 ÄĲá»ĥ

0.25

ÑĤÐ¾Ð±Ñĭ

0.25

 Ð§ÑĤÐ¾Ð±Ñĭ

0.25

Activations Density 0.141%