INDEX

Explanations

be or become

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

Å¼

-0.10

ãĤ

-0.10

zik

-0.09

 Farrell

-0.09

çĦ¡ãģĹãģ

-0.09

 Mature

-0.08

 competing

-0.08

eniable

-0.08

 firm

-0.08

åŃĺæ¡£

-0.08

POSITIVE LOGITS

 anything

0.16

anything

0.12

 something

0.12

è¿«

0.11

 alike

0.10

simp

0.10

 ÚĨÛĮØ²ÛĮ

0.10

haf

0.10

omething

0.09

haft

0.09

Activations Density 0.116%