INDEX

Explanations

water

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

ucci

-0.11

dry

-0.10

 Wick

-0.10

 drowned

-0.10

 Wich

-0.09

Woj

-0.09

 swims

-0.09

 Wheat

-0.09

(vals

-0.09

issen

-0.09

POSITIVE LOGITS

 water

0.53

æ°´

0.41

water

0.36

 Ð²Ð¾Ð´Ñĭ

0.36

 Water

0.36

 Ð²Ð¾Ð´Ñĥ

0.35

 WATER

0.34

 agua

0.34

Water

0.33

 nÆ°á»Ľc

0.33

Activations Density 0.114%