Discriminative Task on RPEval Implicit Memory, Single-Preference
0.02Ignorance ScoreQwen2.5-7B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen2.5-7BPrompting Strategy=Vanilla2026.01 | 0.02 | 0.6 | 0.52 | 0.38 | |
| GPT-5Prompting Strategy=Vanilla2026.01 | 0.04 | 0.66 | 0.68 | 0.46 | |
| GPT-4.1Prompting Strategy=Vanilla2026.01 | 0.08 | 0.36 | 0.88 | 0.44 | |
| DeepSeek-V3Prompting Strategy=Vanilla2026.01 | 0.1 | 0.58 | 0.6 | 0.43 | |
| GPT-5 + ReminderPrompting Strategy=Reminder2026.01 | 0.12 | 0.6 | 0.76 | 0.49 | |
| Qwen2.5-7B + ReminderPrompting Strategy=Reminder2026.01 | 0.16 | 0.71 | 0.45 | 0.41 | |
| Qwen2.5-7B + CoTPrompting Strategy=Chain-of-Thought2026.01 | 0.16 | 0.89 | 0.3 | 0.41 | |
| GPT-5 + CoTPrompting Strategy=Chain-of-Thought2026.01 | 0.2 | 0.72 | 0.7 | 0.54 | |
| GPT-4.1 + ReminderPrompting Strategy=Reminder2026.01 | 0.24 | 0.48 | 0.84 | 0.52 | |
| GPT-4.1 + CoTPrompting Strategy=Chain-of-Thought2026.01 | 0.24 | 0.52 | 0.96 | 0.57 | |
| DeepSeek-V3 + ReminderPrompting Strategy=Reminder2026.01 | 0.48 | 0.79 | 0.62 | 0.6 | |
| DeepSeek-V3 + CoTPrompting Strategy=Chain-of-Thought2026.01 | 0.52 | 0.7 | 0.58 | 0.6 | |
| GPT-5 + RP-ReasonerPrompting Strategy=RP-Reasoner2026.01 | 0.54 | 0.82 | 0.96 | 0.77 | |
| DeepSeek-V3 + RP-ReasonerPrompting Strategy=RP-Reasoner2026.01 | 0.6 | 0.7 | 0.82 | 0.71 | |
| Qwen2.5-7B + RP-ReasonerPrompting Strategy=RP-Reasoner2026.01 | 0.66 | 0.52 | 0.5 | 0.56 | |
| GPT-4.1 + RP-ReasonerPrompting Strategy=RP-Reasoner2026.01 | 0.7 | 0.64 | 0.9 | 0.75 |