INDEX

Explanations

scientist who works

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

aylor

-0.10

 Manuals

-0.09

 Dipl

-0.09

 Bast

-0.09

 Nurs

-0.09

psych

-0.09

 è¡

-0.08

 Skill

-0.08

ignal

-0.08

POSITIVE LOGITS

lab

0.28

 research

0.24

 scientists

0.23

 scientist

0.22

 laboratory

0.20

Lab

0.20

 scientific

0.18

çłĶç©¶

0.18

research

0.18

 researchers

0.18

Activations Density 0.158%

No Known Activations