INDEX

Explanations

only or alone

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

 collectively

-0.12

imers

-0.10

 orally

-0.09

 yourselves

-0.09

 professionally

-0.09

 unst

-0.09

.jp

-0.08

unp

-0.08

 Fathers

-0.08

ican

-0.08

POSITIVE LOGITS

 solo

0.64

 Solo

0.50

 alone

0.46

Solo

0.46

 lone

0.37

alone

0.36

 Alone

0.35

 sola

0.34

 Ð¾Ð´Ð¸

0.33

-alone

0.33

Activations Density 0.148%