EIBench
Benchmarks
TaskDatasetSOTA resultTrendResults
EIBench Basic 1.0 (test)
69.68Happy Score (LLaMA-3)
19
EIBench (test)
43.8CHARM Score
15
EIBench
39.27Recall (LLaMA-3 Criteria)
13
EIBench DeepSeek-v4-pro simulator
23.7Score
4
EIBench Qwen3-Max simulator
28.7Score
4