INDEX

Explanations

introducing details or examples

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

endregion

-0.09

imon

-0.09

iverz

-0.08

 Ones

-0.08

eus

-0.08

 Exercises

-0.08

 Richt

-0.08

astes

-0.08

 simply

-0.08

POSITIVE LOGITS

 including

0.29

åĮħæĭ¬

0.26

including

0.23

 includes

0.21

 includ

0.20

 include

0.20

 vÄįetnÄĽ

0.20

includes

0.17

Including

0.17

 INCLUDING

0.17

Activations Density 0.326%