INDEX

Explanations

levels of substances

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

 Gross

-0.11

 Troll

-0.10

 hormonal

-0.09

Î»Î¹

-0.09

-g

-0.09

çİī

-0.09

 Shapiro

-0.09

bero

-0.09

 respiratory

-0.09

 gross

-0.08

POSITIVE LOGITS

 levels

0.50

levels

0.39

 Levels

0.37

Levels

0.34

_levels

0.28

 level

0.28

 nive

0.26

.levels

0.26

(levels

0.22

 ÑĥÑĢÐ¾Ð²

0.21

Activations Density 0.056%