INDEX

Explanations

possessive pronoun and location

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

 pouÅ¾it

-0.10

 Classroom

-0.10

 classrooms

-0.10

 classroom

-0.10

 suburbs

-0.09

 pinnacle

-0.09

Rooms

-0.09

imos

-0.09

 nightlife

-0.09

 bedrooms

-0.09

POSITIVE LOGITS

 expense

0.21

 helm

0.17

 glance

0.17

expense

0.15

 mercy

0.15

 Expense

0.14

 discretion

0.13

 request

0.13

 location

0.13

 desk

0.13

Activations Density 0.065%