INDEX

Explanations

development of content filtration

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

azy

-0.10

 Cunning

-0.09

 Lara

-0.08

uhl

-0.08

.mob

-0.08

Wah

-0.08

elt

-0.08

bou

-0.08

iy

-0.08

POSITIVE LOGITS

 testing

0.13

 test

0.12

AI

0.12

 development

0.11

afety

0.10

 flag

0.10

.ai

0.10

 safety

0.10

æµĭè¯ķ

0.10

_AI

0.09

Activations Density 0.078%