INDEX

Explanations

or health

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

 bour

-0.10

 utens

-0.09

 Ð·Ð°ÑıÐ²

-0.09

 inse

-0.09

 maturity

-0.09

ÑģÐ¸Ð»ÑĮ

-0.08

Ð¾Ð¶Ð´ÐµÐ½Ð¸Ðµ

-0.08

YC

-0.08

bero

-0.08

 domest

-0.08

POSITIVE LOGITS

 stab

0.15

 degree

0.13

 level

0.12

ÌĨ

0.11

 indicators

0.11

 ÑįÑĦÑĦÐµÐº

0.10

degree

0.10

vais

0.10

 quality

0.10

 rating

0.09

Activations Density 0.078%