Question Answering on OpenBookQA (Mean per-step regret)
0.157Mean Per-Step RegretSimpl
Evaluation Results
| Method | Links | |
|---|---|---|
| SimplStrategy Category=Static Prompts2026.02 | 0.157 | |
| ParaStrategy Category=Static Prompts2026.02 | 0.159 | |
| TSStrategy Category=Non-Contextual2026.02 | 0.159 | |
| No-Rewrite (NoRw)Strategy Category=Base2026.02 | 0.16 | |
| FTPLStrategy Category=Non-Contextual2026.02 | 0.169 | |
| ϵ-FTRLStrategy Category=Non-Contextual2026.02 | 0.177 | |
| TSStrategy Category=Contextual Linear2026.02 | 0.188 | |
| LinUCB+KLStrategy Category=Contextual Linear2026.02 | 0.198 | |
| LinUCBStrategy Category=Contextual Linear2026.02 | 0.2 | |
| DisambStrategy Category=Static Prompts2026.02 | 0.218 | |
| LinFTPLStrategy Category=Contextual Linear2026.02 | 0.221 | |
| EXP3Strategy Category=Non-Contextual2026.02 | 0.223 | |
| ClarifyStrategy Category=Static Prompts2026.02 | 0.228 | |
| LinEXP3Strategy Category=Contextual Linear2026.02 | 0.243 | |
| ExpandStrategy Category=Static Prompts2026.02 | 0.341 |