INDEX

Explanations

ratios and scores

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

 Chim

-0.09

caff

-0.09

outh

-0.09

$MESS

-0.09

çļĦä¸Ģ

-0.08

/xhtml

-0.08

ä¸¶

-0.08

odash

-0.08

ijd

-0.08

ensi

-0.08

POSITIVE LOGITS

 total

0.13

total

0.12

 possibile

0.12

 possible

0.12

possible

0.11

 tá»ķng

0.11

 posible

0.11

ths

0.10

 Possible

0.10

Activations Density 0.061%