INDEX

Explanations

help and advice

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

of

-0.12

 secara

-0.10

 alias

-0.10

-0.09

anki

-0.09

 Shel

-0.09

 unten

-0.08

 prosper

-0.08

neck

-0.08

 backward

-0.08

POSITIVE LOGITS

 decent

0.09

ãģĨãģ¡

0.09

 Äĳá»ĥ

0.09

ÐļÐ¢

0.09

oyo

0.09

UNUSED

0.08

 nerd

0.08

 Ø¬ÙĩØª

0.08

itzer

0.08

DevExpress

0.08

Activations Density 0.145%