INDEX

Explanations

organizing labor and workers

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

 affluent

-0.10

 witch

-0.09

asco

-0.08

ÑĥÐ»Ñİ

-0.08

 High

-0.08

 outward

-0.08

 elders

-0.08

_RG

-0.08

ussen

-0.08

/^\

-0.08

POSITIVE LOGITS

 working

0.41

 Working

0.33

working

0.32

Working

0.30

 ÑĢÐ°Ð±Ð¾Ñĩ

0.29

-working

0.28

 workers

0.28

 worker

0.26

 lower

0.24

 class

0.24

Activations Density 0.069%