INDEX

Explanations

customer support or service

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

 private

-0.10

èĪį

-0.09

 Fatal

-0.09

 Technician

-0.08

 Solic

-0.08

 Industry

-0.08

 aide

-0.08

 relie

-0.08

 Stephens

-0.08

 Chief

-0.08

POSITIVE LOGITS

 support

0.27

 customer

0.25

 Support

0.24

hel

0.22

 Customer

0.22

Customer

0.18

Support

0.18

support

0.18

customer

0.17

_support

0.17

Activations Density 0.128%