INDEX

Explanations

smooth functioning and operation

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

à¸«à¸¡à¸Ķ

-0.09

ëį°ìĿ´íĬ¸

-0.09

arness

-0.08

 Nichols

-0.08

 unused

-0.08

?action

-0.08

 Gors

-0.08

auss

-0.08

oley

-0.08

 elabor

-0.08

POSITIVE LOGITS

 smooth

0.34

 functioning

0.33

smooth

0.29

 Smooth

0.28

 funcion

0.27

Smooth

0.26

 functionality

0.23

 operation

0.23

 running

0.23

 function

0.22

Activations Density 0.125%