INDEX

Explanations

cultivation and cultural meanings

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

ity

-0.11

 Sherman

-0.10

hou

-0.10

ertools

-0.10

alties

-0.09

Ð»Ð¾Ðº

-0.09

erb

-0.09

hill

-0.09

RetVal

-0.09

idUser

-0.09

POSITIVE LOGITS

urally

0.20

ures

0.17

URAL

0.17

ture

0.16

ured

0.16

ura

0.15

ivation

0.15

ovnÃŃ

0.14

prit

0.13

utral

0.13

Activations Density 0.019%