INDEX

Explanations

first, launch, project, know

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

eler

-0.10

 proto

-0.10

ãĤ©

-0.09

:animated

-0.09

ÂłBf

-0.08

elez

-0.08

 Secondly

-0.08

_ASSUME

-0.08

 OnTrigger

-0.08

 Brig

-0.08

POSITIVE LOGITS

 first

0.49

åħĪ

0.40

first

0.40

ãģ¾ãģļ

0.37

 ë¨¼ìłĢ

0.35

Ð½Ð°ÑĩÐ°Ð»Ð°

0.33

 primero

0.33

 åħĪ

0.33

 primeiro

0.30

 Ã¶nce

0.28

Activations Density 0.190%