INDEX

Explanations

gain valuable experience

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

arel

-0.10

 inherited

-0.09

kor

-0.09

 investments

-0.09

 cons

-0.09

 Roch

-0.09

uber

-0.09

 Johns

-0.09

uat

-0.09

POSITIVE LOGITS

 experience

0.34

 gain

0.33

 gaining

0.30

 Gain

0.30

 gained

0.30

gain

0.27

Gain

0.27

 exposure

0.27

 Experience

0.27

experience

0.26

Activations Density 0.052%