INDEX

Explanations

the cell

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

(EXPR

-0.10

opak

-0.10

ERRU

-0.09

 Ellison

-0.08

luÅ¾

-0.08

Â©Â©

-0.08

Ã¼ven

-0.08

ogs

-0.08

IXEL

-0.08

utex

-0.08

POSITIVE LOGITS

redirectTo

0.09

ItemSelected

0.08

 maxim

0.08

-Encoding

0.08

FileSync

0.08

0.07

SetActive

0.07

Nin

0.07

Tun

0.07

"s

0.07

Activations Density 0.279%