INDEX

Explanations

the largest

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

.onStart

-0.09

udge

-0.09

inel

-0.08

 Slug

-0.08

APIView

-0.08

amo

-0.08

rieg

-0.08

ince

-0.08

iar

-0.08

.infinity

-0.08

POSITIVE LOGITS

 leading

0.16

larg

0.14

 largest

0.14

 players

0.14

 larg

0.12

 player

0.12

 region

0.12

players

0.11

leading

0.11

 Players

0.11

Activations Density 0.049%