INDEX

Explanations

German cities and postal codes

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

 Oakland

-0.12

Bav

-0.11

abay

-0.11

 Hampshire

-0.11

 Austria

-0.11

 Boston

-0.11

 Austrian

-0.11

bye

-0.11

Boston

-0.11

 whale

-0.10

POSITIVE LOGITS

 Cologne

0.21

Rh

0.19

Rhe

0.18

Bon

0.17

NR

0.17

 Lever

0.15

 Rating

0.13

 kÃ¶ln

0.13

ologne

0.13

NR

0.13

Activations Density 0.044%