INDEX

Explanations

who need or prefer

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

ille

-0.10

ichen

-0.09

Nug

-0.09

 ÐĴÐ»Ð°Ð´Ð¸

-0.09

borne

-0.09

 gonna

-0.08

FO

-0.08

ichten

-0.08

 industrial

-0.08

 chalk

-0.08

POSITIVE LOGITS

 need

0.19

éľĢè¦ģ

0.18

 cáº§n

0.16

 needs

0.16

 want

0.16

 muá»ĳn

0.15

 wants

0.14

 éľ

0.14

 prefer

0.13

need

0.13

Activations Density 0.057%