INDEX

Explanations

telling and offering information

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

ãĤ

-0.10

iband

-0.09

kees

-0.09

/request

-0.08

ask

-0.08

signals

-0.08

 REQUEST

-0.08

emd

-0.08

 understanding

-0.08

 Maiden

-0.08

POSITIVE LOGITS

 tell

0.61

åĳĬè¯ī

0.54

 tells

0.52

 telling

0.50

tell

0.46

 told

0.46

 Tell

0.45

Tell

0.43

 Tells

0.38

.tell

0.37

Activations Density 0.201%