Genetic Circuit Design on Literature-91
44.9Task Success Rate (TSR)Qwen3-8B RLVF-H-C
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-8B RLVF-H-CTraining Protocol=GenCircuit-RL (Hierarchical-Curriculum)2026.05 | 44.9 | |
| Gemma-3-12B RLVF-H-CTraining Protocol=GenCircuit-RL (Hierarchical-Curriculum)2026.05 | 39 | |
| Qwen3-8B RLVF-HIERTraining Protocol=RLVF-Hierarchical2026.05 | 37.8 | |
| Llama-3.1-8B RLVF-H-CTraining Protocol=GenCircuit-RL (Hierarchical-Curriculum)2026.05 | 37.1 | |
| Qwen3-8B RLVF-BINTraining Protocol=RLVF-Binary2026.05 | 31.1 | |
| Qwen3-8B SFTTraining Protocol=Supervised fine-tuning2026.05 | 27.3 | |
| Gemma-3-12B SFTTraining Protocol=Supervised fine-tuning2026.05 | 23.4 | |
| Llama-3.1-8B SFTTraining Protocol=Supervised fine-tuning2026.05 | 22.4 | |
| OPUS 4.5Few-shot Setting=5-shot2026.05 | 21.7 | |
| OPUS 4.5Few-shot Setting=0-shot2026.05 | 14.8 | |
| Qwen3-8B BASEFew-shot Setting=0-shot2026.05 | 4.4 |