INDEX

Explanations

explaining and in addition

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

 physical

-0.18

 Physical

-0.17

 physically

-0.16

Physical

-0.16

physical

-0.15

_physical

-0.13

 fizik

-0.13

 fÃŃs

-0.13

èº«ä½ĵ

-0.12

çī©çĲĨ

-0.12

POSITIVE LOGITS

æ³ī

0.10

 chemical

0.10

 facial

0.10

Smarty

0.10

 electronics

0.09

 vocalist

0.09

 Electronics

0.09

 electrical

0.09

 vocal

0.09

akers

0.09

Activations Density 0.176%