INDEX

Explanations

is located in

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

 indeb

-0.10

YRO

-0.09

ihar

-0.08

 Jord

-0.08

ercul

-0.08

onso

-0.08

 emergence

-0.08

æľ¬å½ĵ

-0.08

usp

-0.08

 reife

-0.08

POSITIVE LOGITS

 befind

0.25

 Ð½Ð°ÑħÐ¾Ð´Ð¸ÑĤÑģÑı

0.22

 exist

0.20

 lies

0.20

 exists

0.18

lie

0.18

 náº±m

0.17

 Î²ÏģÎ¯

0.17

à¸ĩà¸Ńà¸¢

0.17

 Ð½Ð°ÑħÐ¾Ð´

0.16

Activations Density 0.201%