INDEX

Explanations

statistical analysis and data manipulation

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

 analytic

-0.11

 intelligence

-0.10

 intel

-0.09

adal

-0.09

 Rewrite

-0.09

mate

-0.09

 pulumi

-0.09

ilight

-0.09

éľĩ

-0.09

ilty

-0.09

POSITIVE LOGITS

wr

0.17

Wr

0.14

 cleaning

0.13

 Cleans

0.13

 explor

0.13

 exploration

0.13

 Expl

0.11

 business

0.11

Wr

0.11

 Exploration

0.11

Activations Density 0.078%