Short-Horizon Analytical Question Answering on Wearable Sensing Datasets (test)
95.6Exact Match (EM)WEQA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| WEQATokens=10,4902026.06 | 95.6 | 9.2 | |
| Multi-AgentTokens=32,3412026.06 | 72 | 157.2 | |
| ReActTokens=41,9022026.06 | 64.8 | 31.3 | |
| LLM-TextGran.=daily, Tokens=72,2342026.06 | 10 | 1,013 | |
| LLM-TextGran.=hourly, Tokens=112,3282026.06 | 6 | 848 | |
| LLM-ImageGran.=hourly, Tokens=3,9772026.06 | 3.2 | 526.4 | |
| LLM-ImageGran.=daily, Tokens=4,6742026.06 | 1.2 | 913.6 | |
| Random2026.06 | 0.4 | 5,345.6 |