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

Novel

Benchmarks

Task NameDataset NameSOTA ResultTrend
Long-form Question AnsweringNovel GraphRAG-Bench
LLM-Acc85.3
20
Object DetectionNovel-114 Average
AP50:9555.1
11
Retrieval-Augmented GenerationNovel
Indexing Time (mins)13
11
Retrieval EfficiencyNovel
Retrieved Tokens22,391
8
Object DetectionNovel-114 Instruments-8
AP58.5
3
Object DetectionNovel-114 Cold Weapons-4
AP74.4
3
Surgical robot end-effector pose estimationNovel unseen configuration (test)
RMSE X (mm)1.37
2
Showing 7 of 7 rows