INDEX

Explanations

closest, best, smallest, latest

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

 remainder

-0.13

 rest

-0.09

_UNS

-0.09

 Morse

-0.09

lier

-0.09

rud

-0.09

 lightning

-0.09

ï¼ı/

-0.08

 Vill

-0.08

POSITIVE LOGITS

æľĢè¿ĳ

0.19

 closest

0.16

fur

0.15

fur

0.15

 highest

0.15

 Ð½Ð°Ð¸Ð±Ð¾Ð»ÐµÐµ

0.15

 closet

0.15

closest

0.14

 ê°Ģìŀ¥

0.14

 lowest

0.14

Activations Density 0.106%