INDEX

Explanations

safety and related situations

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

Negative Logits

 homework

-0.10

 Statistics

-0.09

 integr

-0.09

 junk

-0.09

 staff

-0.09

 Spare

-0.09

 statistics

-0.09

Vol

-0.08

 engineers

-0.08

 Junk

-0.08

POSITIVE LOGITS

 safety

0.25

 Safety

0.21

å®īåħ¨

0.20

Safety

0.18

afety

0.17

 safe

0.17

 safely

0.17

 unsafe

0.16

 Ð±ÐµÐ·Ð¾Ð¿Ð°Ñģ

0.15

 safest

0.15

Activations Density 0.090%

No Known Activations