INDEX

Explanations

diligently studying or learning

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

 Instructor

-0.10

 instructor

-0.10

 mezun

-0.09

STRUCTOR

-0.09

 Ã¶ÄŁretmen

-0.09

 graduates

-0.09

Faculty

-0.09

 grads

-0.08

/Instruction

-0.08

EqualTo

-0.08

POSITIVE LOGITS

 study

0.48

 learning

0.42

 studying

0.41

 Study

0.39

åŃ¦ä¹ł

0.39

study

0.38

 studies

0.37

Study

0.35

 Ð¸Ð·ÑĥÑĩ

0.33

 estud

0.33

Activations Density 0.240%