INDEX

Explanations

additional context or information

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

 additional

-0.11

 additions

-0.09

oti

-0.09

eur

-0.09

posium

-0.09

 redirectTo

-0.09

 adicion

-0.09

kek

-0.09

 ayrÄ±ca

-0.08

POSITIVE LOGITS

0.15

ities

0.14

nal

0.14

mente

0.14

/new

0.13

ity

0.12

à¸¡à¹Ģà¸ķ

0.12

ìłģìĿ¸

0.12

-large

0.12

CTION

0.12

Activations Density 0.016%