Multiple Choice Question Answering on OpenBookQA 0-shot (test)
33.2AccuracyQwen1.5-MoE-A2.7B with training-free loop wrapper
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen1.5-MoE-A2.7B with training-free loop wrapperModel=Qwen1.5-MoE-A2.7B, Evaluation Protocol=0-shot, Loop Setting=Loop (Ours), Backbone Architecture=MoE, Iteration Mode=layer-mode, Integration Recipe=3-stage Runge–Kutta2026.05 | 33.2 | 1.6 | |
| Moonlight-16B-A3B with training-free loop wrapperModel=Moonlight-16B-A3B, Evaluation Protocol=0-shot, Loop Setting=Loop (Ours), Backbone Architecture=MoE, Iteration Mode=layer-mode, Integration Recipe=3-stage Runge–Kutta2026.05 | 32.8 | 1.2 | |
| Qwen1.5-MoE-A2.7BModel=Qwen1.5-MoE-A2.7B, Evaluation Protocol=0-shot, Loop Setting=Base, Backbone Architecture=MoE2026.05 | 31.6 | — | |
| Moonlight-16B-A3BModel=Moonlight-16B-A3B, Evaluation Protocol=0-shot, Loop Setting=Base, Backbone Architecture=MoE2026.05 | 31.6 | — |