INDEX

Explanations

in the lungs

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

Ã³c

-0.10

_Tis

-0.10

.Dictionary

-0.10

 DÃ¼n

-0.09

ufen

-0.09

elight

-0.09

isContained

-0.09

cobra

-0.09

vironments

-0.09

iffany

-0.09

POSITIVE LOGITS

 otherwise

0.12

imit

0.11

 OTHERWISE

0.10

otherwise

0.10

iki

0.09

 Otherwise

0.09

arak

0.09

 necessity

0.09

 Levy

0.08

 strat

0.08

Activations Density 0.181%