INDEX

Explanations

identifying specific enzymes and their roles

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

 Vine

-0.09

anford

-0.09

 peptides

-0.09

 Scha

-0.09

 antig

-0.09

 Foley

-0.09

patible

-0.09

GenerationStrategy

-0.09

ControlItem

-0.09

avo

-0.09

POSITIVE LOGITS

 enzymes

0.21

ase

0.20

 activity

0.19

ases

0.18

 enzyme

0.17

activity

0.16

 catalyst

0.15

 activities

0.15

 Activity

0.15

rase

0.14

Activations Density 0.055%