Student Writing Feedback Generation on SEFORA (full corpus)
31.7F1 ScoreQwen_72B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen_72Bshot-count=few-shot, rubric=included, guidance=guided, prompt template=constrained2026.06 | 31.7 | — | — | |
| GPT-5.1shot-count=zero-shot, rubric=included, guidance=guided, prompt template=constrained2026.06 | 31.6 | 37.6 | 27.3 | |
| Qwen_72Bshot-count=zero-shot, rubric=included, guidance=guided, prompt template=constrained2026.06 | 31.2 | 37.4 | 26.8 | |
| GPT-5.1shot-count=few-shot, rubric=included, guidance=guided, prompt template=constrained2026.06 | 30.7 | — | — | |
| GPT-5.1shot-count=zero-shot, rubric=included, guidance=unguided, prompt template=constrained2026.06 | 29.6 | — | — | |
| GPT-5.1shot-count=zero-shot, rubric=no rubrics, guidance=guided, prompt template=constrained2026.06 | 28.8 | — | — | |
| Qwen_72Bshot-count=zero-shot, rubric=no rubrics, guidance=guided, prompt template=constrained2026.06 | 28.5 | — | — | |
| Mistral_7Bshot-count=zero-shot, rubric=no rubrics, guidance=guided, prompt template=constrained2026.06 | 27.8 | — | — | |
| Mistral_7Bshot-count=zero-shot, rubric=included, guidance=guided, prompt template=constrained2026.06 | 27.7 | 28.8 | 26.7 | |
| Llama_8Bshot-count=few-shot, rubric=included, guidance=guided, prompt template=constrained2026.06 | 27.1 | — | — | |
| Qwen_72Bshot-count=zero-shot, rubric=included, guidance=unguided, prompt template=constrained2026.06 | 27.1 | — | — | |
| Llama_8Bshot-count=zero-shot, rubric=no rubrics, guidance=guided, prompt template=constrained2026.06 | 26.8 | — | — | |
| Mistral_7Bshot-count=few-shot, rubric=included, guidance=guided, prompt template=constrained2026.06 | 25.8 | — | — | |
| Llama_8Bshot-count=zero-shot, rubric=included, guidance=guided, prompt template=constrained2026.06 | 25.4 | 27.5 | 23.6 | |
| Mistral_7Bshot-count=zero-shot, rubric=included, guidance=unguided, prompt template=constrained2026.06 | 25.1 | — | — | |
| Llama_8Bshot-count=zero-shot, rubric=included, guidance=unguided, prompt template=constrained2026.06 | 20.3 | — | — | |
| Qwen_72Bshot-count=zero-shot, rubric=included, guidance=guided, prompt template=unconstrained2026.06 | 17.9 | — | — | |
| Mistral_7Bshot-count=zero-shot, rubric=included, guidance=guided, prompt template=unconstrained2026.06 | 15.9 | — | — | |
| GPT-5.1shot-count=zero-shot, rubric=included, guidance=guided, prompt template=unconstrained2026.06 | 13.6 | — | — | |
| Llama_8Bshot-count=zero-shot, rubric=included, guidance=guided, prompt template=unconstrained2026.06 | 11.3 | — | — |