INDEX

Explanations

leadership drive and strength

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

 humility

-0.11

 aver

-0.11

removeAttr

-0.09

 Grund

-0.09

 undes

-0.09

hum

-0.09

oyo

-0.09

:|:

-0.09

 orchestrated

-0.09

itter

-0.09

POSITIVE LOGITS

 leadership

0.20

 Leadership

0.15

 leader

0.15

 drive

0.14

 DRIVE

0.12

leaders

0.12

 Assert

0.12

 Leader

0.12

 assert

0.12

leader

0.12

Activations Density 0.109%