INDEX

Explanations

pronouns referring to self

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

 same

-0.11

ï¼«

-0.10

Same

-0.10

alone

-0.10

 SAME

-0.09

 âĦ

-0.09

same

-0.09

Own

-0.09

ÑĥÑİÑĩÐ¸

-0.09

ï½ŀï½ŀ

-0.09

POSITIVE LOGITS

 himself

0.28

 itself

0.28

 herself

0.27

 themselves

0.26

 zich

0.25

 ÑģÐµÐ±Ñı

0.23

 siÄĻ

0.23

 mÃ¬nh

0.22

èĩªå·±

0.22

 sich

0.21

Activations Density 0.056%