INDEX

Explanations

treatment with antibiotics

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

 çĻ

-0.11

 cancer

-0.11

 Cancer

-0.11

 revers

-0.10

 inoc

-0.09

Rao

-0.09

uet

-0.09

 reversible

-0.09

 cancers

-0.09

vas

-0.09

POSITIVE LOGITS

 culture

0.15

 antibiotics

0.15

pen

0.14

 stre

0.14

 Culture

0.13

culture

0.13

 empirical

0.13

Pen

0.13

 antibiotic

0.13

 treatment

0.13

Activations Density 0.032%