INDEX

Explanations

animal attributes and qualities

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

æĺ¯ä¸ª

-0.10

.ct

-0.10

psy

-0.09

 Shark

-0.09

Ð¼ÐµÑĢÐ¸Ðº

-0.09

_PTR

-0.09

acha

-0.09

 carniv

-0.09

 ÙħØµ

-0.09

iddy

-0.09

POSITIVE LOGITS

 fasc

0.18

 intelligence

0.17

 Fasc

0.17

 fascinating

0.16

 beauty

0.15

maj

0.14

maj

0.13

 grace

0.13

 Intelligence

0.12

Maj

0.12

Activations Density 0.160%