Expert Evaluation
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Expert Evaluation Feedback Quality
4.01Accuracy
5
Expert Evaluation
3.97Specificity
4
Expert Evaluation Biotech Domain (test)
91.7Novelty Winrate vs Qwen-14B
1
Expert Evaluation Education Domain (test)
80Novelty Winrate vs Qwen
1
Expert Evaluation Law Domain (test)
69.2Novelty (Winrate vs Qwen)
1