INDEX

Explanations

introductions like here, aqui, hier

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

earer

-0.10

 Cong

-0.08

arent

-0.08

ToUpper

-0.08

Cong

-0.08

descricao

-0.08

ìĥģìĿĦ

-0.08

ellan

-0.08

HIR

-0.07

EEEE

-0.07

POSITIVE LOGITS

 here

0.55

 Here

0.45

Here

0.43

here

0.41

 aquÃŃ

0.36

 below

0.34

 aqui

0.34

_here

0.33

 HERE

0.33

è¿ĻéĩĮ

0.33

Activations Density 0.209%