INDEX

Explanations

i offering assistance or help

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

libs

-0.09

uci

-0.09

 Yourself

-0.09

 Buchanan

-0.09

ully

-0.08

 ragaz

-0.08

isy

-0.08

 Guth

-0.08

pees

-0.08

OWN

-0.08

POSITIVE LOGITS

 assistance

0.21

 assist

0.19

 assists

0.17

 help

0.17

 Assistance

0.16

 assisting

0.15

 helped

0.15

åĬ©

0.15

 giÃºp

0.14

 helping

0.14

Activations Density 0.056%