INDEX

Explanations

front or before in multiple languages

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

 Ð¿ÑĢÐµÐ¶Ð´Ðµ

-0.11

PRI

-0.10

 Prior

-0.10

 overhead

-0.09

athers

-0.09

antic

-0.09

 buzz

-0.09

 prior

-0.09

ereo

-0.08

Ned

-0.08

POSITIVE LOGITS

 devant

0.54

 front

0.47

 frente

0.42

front

0.38

 ante

0.30

 ìķŀ

0.30

 Ð¿ÐµÑĢÐµÐ´

0.30

åīį

0.30

_front

0.29

 Ã¶nÃ¼nde

0.28

Activations Density 0.134%