Post-argument rating regression on Anthropic dataset September 2024
0.67MSEMS-PS-MLP
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MS-PS-MLPPrompting Strategy=MS-PS, Model Architecture=MLP2026.01 | 0.67 | 0.8185 | 0.6339 | 0.811 | 46.2 | |
| Baseline-2-o3LLM Backbone=OpenAI-o3, Prompting Strategy=Baseline-2 (refined)2026.01 | 0.9882 | 0.9941 | 0.731 | 0.721 | 37.9 | |
| Baseline-1-o3LLM Backbone=OpenAI-o3, Prompting Strategy=Baseline-1 (single-prompt)2026.01 | 1.0068 | 1.0034 | 0.7394 | 0.716 | 37.4 | |
| Baseline-2-GemmaLLM Backbone=Gemma, Prompting Strategy=Baseline-2 (refined)2026.01 | 3.5854 | 1.8935 | 1.5347 | -0.011 | 20.1 | |
| Baseline-1-GemmaLLM Backbone=Gemma, Prompting Strategy=Baseline-1 (single-prompt)2026.01 | 4.3553 | 2.0869 | 1.6988 | -0.229 | 18.4 |