INDEX

Explanations

throughout duration or engagement

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

.scalablytyped

-0.12

atus

-0.09

 Overnight

-0.09

 ifndef

-0.08

 recently

-0.08

ault

-0.08

 Recently

-0.08

æľĢè¿ĳ

-0.08

 Everyday

-0.08

_lens

-0.08

POSITIVE LOGITS

 throughout

0.93

 Throughout

0.73

Throughout

0.65

 suá»ĳt

0.44

 during

0.38

à¸ķà¸¥à¸Ńà¸Ķ

0.38

 Ð¿ÑĢÐ¾ÑĤÑıÐ³Ð¾Ð¼

0.35

 durante

0.32

during

0.30

 bÄĽhem

0.28

Activations Density 0.196%