INDEX

Explanations

ethics and relationships

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

 âĢ¦

-0.11

ilden

-0.09

...

-0.09

 âĢª

-0.09

..

-0.09

åĢ

-0.09

iffe

-0.08

&eacute

-0.08

 _______,

-0.08

POSITIVE LOGITS

\\n

0.10

\\n

0.10

ogs

0.09

 ********************************************************

0.09

etc

0.09

ï¼Į\n

0.08

 noqa

0.08

bsp

0.08

 âĨĴ\n\n

0.08

IOD

0.08

Activations Density 0.183%