Automated Short Answer Scoring on ASAP-SAS (test)
0.95Score Q1Human Scorers
Evaluation Results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Human ScorersProtocol=Human Baseline2025.11 | 0.95 | 0.93 | 0.77 | 0.75 | 0.95 | 0.93 | 0.96 | 0.86 | 0.84 | 0.87 | 0.88 | 0.93 | 0.79 | 0.91 | |
| Llama-3.1-8b (Instruct)Protocol=Supervised Fine-tuned, Features=w/ rubric2025.11 | 0.87 | 0.8 | 0.68 | 0.77 | 0.85 | 0.8 | 0.72 | 0.65 | 0.84 | 0.79 | 0.78 | 0.82 | 0.76 | 0.68 | |
| DeBERTaProtocol=Supervised Fine-tuned2025.11 | 0.85 | 0.86 | 0.66 | 0.7 | 0.81 | 0.83 | 0.71 | 0.64 | 0.79 | 0.71 | 0.76 | 0.81 | 0.72 | 0.67 | |
| Llama-3.1-8b (Instruct)Protocol=Supervised Fine-tuned2025.11 | 0.84 | 0.73 | 0.72 | 0.71 | 0.82 | 0.81 | 0.71 | 0.66 | 0.82 | 0.75 | 0.76 | 0.79 | 0.75 | 0.69 | |
| Llama-3.1-8bProtocol=Supervised Fine-tuned2025.11 | 0.83 | 0.75 | 0.7 | 0.77 | 0.82 | 0.84 | 0.68 | 0.65 | 0.82 | 0.74 | 0.76 | 0.8 | 0.76 | 0.67 | |
| ANALYTICSCOREBackbone=GPT-4.1-mini, Interpretable=true2025.11 | 0.8 | 0.86 | 0.64 | 0.59 | 0.79 | 0.78 | 0.61 | 0.59 | 0.8 | 0.68 | 0.72 | 0.78 | 0.68 | 0.6 | |
| ANALYTICSCOREConfig=Distillation, Interpretable=true2025.11 | 0.8 | 0.82 | 0.68 | 0.59 | 0.81 | 0.76 | 0.62 | 0.59 | 0.78 | 0.64 | 0.71 | 0.77 | 0.68 | 0.6 | |
| BERTProtocol=Supervised Fine-tuned2025.11 | 0.8 | 0.8 | 0.7 | 0.7 | 0.8 | 0.81 | 0.69 | 0.68 | 0.84 | 0.71 | 0.75 | 0.79 | 0.74 | 0.69 | |
| GPT-4.1Protocol=Few-Shot2025.11 | 0.69 | 0.65 | 0.61 | 0.65 | 0.72 | 0.61 | 0.34 | 0.57 | 0.76 | 0.69 | 0.63 | 0.67 | 0.68 | 0.45 | |
| AutoSASProtocol=Baseline2025.11 | 0.68 | 0.47 | 0.57 | 0.61 | 0.5 | 0.54 | 0.37 | 0.44 | 0.77 | 0.68 | 0.56 | 0.57 | 0.65 | 0.41 | |
| NAMProtocol=Baseline, Interpretable=true2025.11 | 0.63 | 0.62 | 0.43 | 0.35 | 0.72 | 0.63 | 0.42 | 0.38 | 0.76 | 0.62 | 0.56 | 0.64 | 0.52 | 0.4 | |
| AsRRNProtocol=Baseline2025.11 | 0.6 | 0.43 | 0.57 | 0.6 | 0.61 | 0.64 | 0.59 | 0.51 | 0.71 | 0.66 | 0.59 | 0.59 | 0.63 | 0.55 | |
| ANALYTICSCOREBackbone=Llama-3.1-8b (Instruct), Interpretable=true2025.11 | 0.57 | 0.57 | 0.59 | 0.56 | 0.69 | 0.47 | 0.52 | 0.45 | 0.74 | 0.6 | 0.58 | 0.58 | 0.63 | 0.48 |