INDEX

Explanations

make informed decisions

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

Am

-0.09

RIX

-0.09

ucas

-0.08

rics

-0.08

rix

-0.08

æķ¬

-0.08

 coats

-0.08

xious

-0.08

 deploy

-0.08

Ger

-0.08

POSITIVE LOGITS

 informed

0.36

 decisions

0.29

 decision

0.19

 choices

0.17

decision

0.16

 Decision

0.15

 rozhodnutÃŃ

0.15

formed

0.15

 wise

0.15

educated

0.14

Activations Density 0.021%