INDEX

Explanations

possessive pronouns + "'s" or "own"

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

Hin

-0.09

Zar

-0.09

ãĥ½

-0.08

ienne

-0.08

Ã³r

-0.08

ichier

-0.08

 facial

-0.08

olds

-0.08

 Facial

-0.08

ex

-0.08

POSITIVE LOGITS

own

0.25

 sights

0.24

OWN

0.16

 prÃ³p

0.16

Own

0.15

 fair

0.14

Own

0.14

 eigenen

0.14

own

0.14

 eyes

0.13

Activations Density 0.051%