Time Series Reasoning on SLEEP QA
0.7255AccTimeMaster
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TimeMasterRL=true2026.02 | 0.7255 | — | |
| Base Model2026.02 | 0.4944 | — | |
| TimeMaster2026.02 | 0.4751 | — | |
| ARTISTmode=+ SFT + RL2026.02 | 0.3663 | 0.1921 | |
| ARTISTRL=true, SFT=true2026.02 | 0.3663 | — | |
| OpenTSLM-4Bmode=+ SFT2026.02 | 0.3537 | 0.1899 | |
| ITFormer-4Bmode=+ SFT2026.02 | 0.3362 | 0.1577 | |
| Llama-3 8Bstatistics=w/o statistics2026.02 | 0.3199 | 0.1409 | |
| ARTISTmode=+ SFT2026.02 | 0.2813 | 0.1794 | |
| ARTISTSFT=true2026.02 | 0.2813 | — | |
| ChatTS-14Bmode=+ SFT2026.02 | 0.2647 | 0.146 | |
| VL-Timefew-shot=true2026.02 | 0.2353 | — | |
| Llama-3 8Bstatistics=w/ statistics2026.02 | 0.2206 | 0.1334 | |
| VL-Time2026.02 | 0.2206 | — | |
| ChatTS-14Bmode=Base Model2026.02 | 0.1765 | 0.1319 | |
| Random Guess2026.02 | 0.1667 | 0.1667 | |
| Qwen3-8Bstatistics=w/o statistics2026.02 | 0.049 | 0.0671 | |
| Qwen3-14Bstatistics=w/o statistics2026.02 | 0.0343 | 0.0573 | |
| Qwen3-14Bstatistics=w/ statistics2026.02 | 0.0343 | 0.0448 | |
| Qwen3-8Bstatistics=w/ statistics2026.02 | 0.0196 | 0.0387 | |
| GPT-5statistics=w/o statistics2026.02 | 0.0049 | 0.0177 | |
| GPT-5statistics=w/ statistics2026.02 | 0.0049 | 0.0149 |