INDEX

Explanations

listing benefits and features

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

iming

-0.10

 heav

-0.09

kn

-0.08

Ard

-0.08

_singular

-0.08

ponder

-0.08

icions

-0.08

otor

-0.08

 profitable

-0.08

loom

-0.08

POSITIVE LOGITS

 improved

0.14

 increased

0.14

 Flex

0.13

 chance

0.12

 convenience

0.12

 reduced

0.12

Flex

0.12

flex

0.12

 better

0.11

 hedge

0.11

Activations Density 0.107%