Genetic Circuit Design on Procedural (test)
72.7Task Success Rate (TSR)Qwen3-8B RLVF-H-C
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-8B RLVF-H-CTraining Protocol=GenCircuit-RL (Hierarchical-Curriculum)2026.05 | 72.7 | |
| Qwen3-8B RLVF-HIERTraining Protocol=RLVF-Hierarchical2026.05 | 67.7 | |
| Gemma-3-12B RLVF-H-CTraining Protocol=GenCircuit-RL (Hierarchical-Curriculum)2026.05 | 66.2 | |
| Llama-3.1-8B RLVF-H-CTraining Protocol=GenCircuit-RL (Hierarchical-Curriculum)2026.05 | 64.1 | |
| Qwen3-8B RLVF-BINTraining Protocol=RLVF-Binary2026.05 | 58.6 | |
| Qwen3-8B SFTTraining Protocol=Supervised fine-tuning2026.05 | 53.9 | |
| Gemma-3-12B SFTTraining Protocol=Supervised fine-tuning2026.05 | 49 | |
| Llama-3.1-8B SFTTraining Protocol=Supervised fine-tuning2026.05 | 47.2 | |
| OPUS 4.5Few-shot Setting=5-shot2026.05 | 41.8 | |
| OPUS 4.5Few-shot Setting=0-shot2026.05 | 30.8 | |
| Qwen3-8B BASEFew-shot Setting=0-shot2026.05 | 10.4 |