INDEX

Explanations

new

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

venture

-0.09

aines

-0.08

OLD

-0.08

 oldValue

-0.08

-generation

-0.08

 inexperienced

-0.08

enuity

-0.08

 rookie

-0.08

Old

-0.07

bery

-0.07

POSITIVE LOGITS

new

0.98

æĸ°çļĦ

0.73

new

0.66

(new

0.59

æĸ°

0.58

 neuen

0.56

 nieuwe

0.56

\tnew

0.56

 neue

0.55

=new

0.55

Activations Density 0.398%