Benchmark
Lo-Hi Drug Discovery
Assesses molecular-property predictions for drug discovery under two task settings, using native ranking evaluations.
42,080items
213subjects
MITlicense
chemistrydomain
medicinedomain
textmodality
item-level responses released
About this data
Dataset name: Lo-Hi Drug Discovery Benchmark
Complete published molecule-level prediction records for the Lo and Hi drug-discovery tasks, with their reference measurements and native aggregate ranking protocol. Individual grades are unavailable.
42,080 items · 1,246,925 recorded attempts. Browse item pages to see all observations. Filters and means apply to the current page.
Loading observation records…