INDEX

Explanations

second list item

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

isia

-0.09

 rounds

-0.09

äºĮäºº

-0.08

 ãģĿãģ®ä»ĸ

-0.08

 Lastly

-0.08

 Rounds

-0.08

ulin

-0.08

hled

-0.08

 Miscellaneous

-0.08

POSITIVE LOGITS

alian

0.11

ï¸ı

0.10

udi

0.09

_navigation

0.09

achs

0.09

hesive

0.08

ffects

0.08

efon

0.08

rol

0.08

xr

0.08

Activations Density 0.159%