INDEX

Explanations

code syntax and structure

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

Negative Logits

{{{

-0.09

 Wolfe

-0.09

anou

-0.08

 Carey

-0.08

 Bethesda

-0.08

 Shea

-0.08

cyn

-0.08

 Harold

-0.08

stride

-0.08

 shorter

-0.08

POSITIVE LOGITS

omik

0.10

endregion

0.10

---</

0.09

 endregion

0.09

ï½Ŀ

0.09

.printStackTrace

0.08

AGR

0.08

-*

0.08

ó

0.08

åĩ½

0.08

Activations Density 0.011%

No Known Activations