Reading Comprehension on DROP (F1 Score)
86.2F1 ScoreScoreFlow
Evaluation Results
| Method | Links | |
|---|---|---|
| ScoreFlow2026.06 | 86.2 | |
| AFlow2026.06 | 83.5 | |
| CoT2026.06 | 83.2 | |
| CoT SC2026.06 | 83.2 | |
| MedPrompt2026.06 | 83 | |
| MetaFlowgeneration_mode=single-inference generation2026.06 | 82.8 | |
| Self Refine2026.06 | 82.5 | |
| IO2026.06 | 81.6 | |
| MultiPersona2026.06 | 81.3 | |
| ADAS2026.06 | 81.3 | |
| openPangu-Embedded RL# Total Params=1B, Training Stage=RL2026.05 | 69.18 | |
| Qwen3# Total Params=1.7B2026.05 | 61.21 | |
| MeZO-LoRAModel=Qwen3-32B, Rank (r)=642026.05 | 58.5 | |
| AR1-ZOModel=Qwen3-32B, Rank (r)=642026.05 | 52.2 | |
| Qwen2.5# Total Params=1.5B2026.05 | 48.44 | |
| openPangu-Embedded KD# Total Params=1B, Training Stage=KD (NPD)2026.05 | 45.69 | |
| Llama3.2# Total Params=1B2026.05 | 45.23 | |
| AR1-ZOModel=Qwen3-1.7B, Rank (r)=642026.05 | 42.9 | |
| MeZO-LoRAModel=Qwen3-1.7B, Rank (r)=642026.05 | 40.1 | |
| Qwen3# Total Params=0.6B2026.05 | 34.69 | |
| FT (Adam)Model=OPT-2.7B2026.05 | 31.4 | |
| FT (Adam)Model=OPT-13B2026.05 | 31.3 | |
| Gemma3# Total Params=1B2026.05 | 30.98 | |
| AR1-ZOModel=OPT-13B, Rank (r)=642026.05 | 30.8 | |
| LOZOModel=OPT-13B2026.05 | 30.6 | |
| MeZO-LoRAModel=OPT-13B, Rank (r)=642026.05 | 30.1 | |
| MiniCPM4# Total Params=0.5B2026.05 | 30.07 | |
| openPangu-Embedded SFT# Total Params=1B, Training Stage=SFT2026.05 | 28.73 | |
| AR1-ZOModel=OPT-2.7B, Rank (r)=642026.05 | 27.8 | |
| MeZO-LoRAModel=OPT-2.7B, Rank (r)=642026.05 | 23.9 | |
| LOZOModel=OPT-2.7B2026.05 | 23.5 | |
| ZO-Alt-NaiveModel=Qwen3-32B2026.05 | 15.1 | |
| ZO-Alt-NaiveModel=Qwen3-1.7B2026.05 | 10.2 | |
| ZO-Alt-NaiveModel=OPT-13B2026.05 | 8.1 | |
| ZO-Alt-NaiveModel=OPT-2.7B2026.05 | 6.3 |