INDEX

Explanations

pumps, produces, provides, allow, responsible

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

 grav

-0.10

rog

-0.10

 land

-0.10

 leads

-0.10

 Leads

-0.09

 Useful

-0.09

 useful

-0.09

 Gaines

-0.09

 activates

-0.09

leads

-0.08

POSITIVE LOGITS

è´Łè´£

0.18

 enable

0.18

 responsible

0.17

 ermÃ¶g

0.16

enable

0.16

 allow

0.15

 plays

0.15

 enables

0.15

ponsible

0.15

 Responsible

0.14

Activations Density 0.171%