INDEX

Explanations

other [noun]

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

 others

-0.14

 other

-0.14

åħ¶ä»ĸ

-0.13

 anderen

-0.12

other

-0.11

others

-0.11

ãģĿãģ®ä»ĸ

-0.11

 OTHER

-0.11

OTHER

-0.10

Others

-0.10

POSITIVE LOGITS

/new

0.16

world

0.15

-than

0.14

wis

0.14

words

0.13

 parts

0.13

ewise

0.13

word

0.12

-wise

0.12

wise

0.12

Activations Density 0.037%