INDEX

Explanations

automatic system operations

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

Bog

-0.10

acker

-0.10

yla

-0.09

ter

-0.09

(Dictionary

-0.09

 either

-0.09

sag

-0.08

deg

-0.08

 Evangel

-0.08

 Hutchinson

-0.08

POSITIVE LOGITS

 autom

0.13

 unless

0.12

 ready

0.12

autom

0.11

 automatically

0.10

 allowing

0.10

unless

0.09

gili

0.09

Autom

0.09

 Unless

0.09

Activations Density 0.056%