INDEX

Explanations

federated learning privacy

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

ëĿ

-0.10

 Diablo

-0.09

 Cutting

-0.09

 Neal

-0.09

Wel

-0.09

nea

-0.09

Dah

-0.09

 æ®

-0.09

licht

-0.08

Neal

-0.08

POSITIVE LOGITS

 privacy

0.22

privacy

0.20

 Privacy

0.19

Privacy

0.18

 sensitive

0.16

_priv

0.14

.priv

0.13

Sensitive

0.13

 differential

0.13

 PRIV

0.12

Activations Density 0.045%