INDEX

Explanations

natural or conversational manner

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

McN

-0.09

 Stevens

-0.09

Äĩi

-0.09

ISO

-0.09

 Hollow

-0.08

ILLE

-0.08

ogne

-0.08

URLException

-0.08

_graphics

-0.08

 Microsystems

-0.08

POSITIVE LOGITS

 natural

0.24

natural

0.18

 Natural

0.17

Natural

0.16

 human

0.16

atural

0.15

lif

0.14

èĩªçĦ¶

0.13

 organic

0.13

 doÄŁal

0.13

Activations Density 0.083%