Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

UltraDomain

Benchmarks

Task NameDataset NameSOTA ResultTrend
Retrieval-Augmented GenerationUltraDomain mix
Win Rate76.2
18
Retrieval-Augmented GenerationUltraDomain agriculture
Win Rate95
18
Question AnsweringUltraDomain Mix 1.0 (512 question-answer pairs)
Exact Match (EM)54.49
10
Question AnsweringUltraDomain Legal 1.0
EM35.16
10
Question AnsweringUltraDomain Computer Science 1.0
EM41.02
10
Question AnsweringUltraDomain Agriculture 1.0
Exact Match (EM)34.96
10
Long-text Question AnsweringUltraDomain
F1 (bio)36.19
10
Conciseness EvaluationUltraDomain
Bio Score100
9
Faithfulness EvaluationUltraDomain
Bio Score100
9
Retrieval-Augmented GenerationUltraDomain Pathology (test)
Comprehension92.35
9
Retrieval-Augmented GenerationUltraDomain Mix (test)
Comprehensiveness91.24
9
Global Question AnsweringUltraDomain Mix
Comprehensiveness76.8
6
Global Question AnsweringUltraDomain Legal
Comprehensiveness60.8
6
Global Question AnsweringUltraDomain Computer Science
Comprehensiveness68.8
6
Global Question AnsweringUltraDomain Agriculture
Comprehensiveness65.6
6
Knowledge Graph Question AnsweringUltraDomain Legal
Comprehensive Score86.19
5
Knowledge Graph Question AnsweringUltraDomain History
Comprehensive87.75
5
Knowledge Graph Question AnsweringUltraDomain Cooking
Comprehensive Score89.25
5
Knowledge Graph Question AnsweringUltraDomain Biology
Comprehensive89.44
5
Knowledge Graph Question AnsweringUltraDomain Agriculture
Comprehensive Score89.93
5
General QAUltraDomain Physics (test)
Comprehensiveness89.51
5
General QAUltraDomain Philosophy (test)
Comprehensive86.78
5
General QAUltraDomain Music (test)
Comprehensiveness87.95
5
General QAUltraDomain Mix (test)
Comprehensive Score87.15
5
General QAUltraDomain Mathematics (test)
Comprehensiveness87.3
5
Showing 25 of 31 rows