Query-Related PII Detection on PII-multi
84.6F1 ScoreNaive w/ Choice
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Naive w/ ChoiceModel=GPT4o2025.02 | 84.6 | 84.6 | |
| Naive w/ ChoiceModel=Qwen2.52025.02 | 80.4 | 80.4 | |
| Naive w/ ChoiceModel=Llama3.12025.02 | 77.5 | 77.5 | |
| Naive w/ ChoiceModel=Qwen2.5-SLM2025.02 | 74.3 | 74.3 | |
| Self-ConsistencyModel=GPT4o2025.02 | 68.5 | 69.2 | |
| Self-CoTModel=GPT4o2025.02 | 67.5 | 68.1 | |
| Auto-CoTModel=Qwen2.5, Prompting Strategy=3-shot2025.02 | 65.7 | 66.5 | |
| Auto-CoTModel=Llama3.1, Prompting Strategy=3-shot2025.02 | 65 | 66.2 | |
| Self-CoTModel=Llama3.12025.02 | 63.8 | 64.3 | |
| PS-CoTModel=Qwen2.52025.02 | 63.6 | 64.3 | |
| Auto-CoTModel=GPT4o, Prompting Strategy=3-shot2025.02 | 62.9 | 64 | |
| Self-CoTModel=Qwen2.52025.02 | 62.6 | 63.2 | |
| PS-CoTModel=Llama3.12025.02 | 62.4 | 63.1 | |
| PS-CoTModel=GPT4o2025.02 | 61.8 | 62 | |
| Self-ConsistencyModel=Qwen2.52025.02 | 61.4 | 62 | |
| NaiveModel=Llama3.12025.02 | 61.1 | 61.4 | |
| Self-ConsistencyModel=Llama3.12025.02 | 60.2 | 60.5 | |
| NaiveModel=GPT4o2025.02 | 60 | 60.2 | |
| NaiveModel=Qwen2.52025.02 | 59.6 | 60.3 | |
| PS-CoTModel=Qwen2.5-SLM2025.02 | 43.1 | 43.6 | |
| NaiveModel=Qwen2.5-SLM2025.02 | 40.5 | 41.3 | |
| Auto-CoTModel=Llama3.1-SLM, Prompting Strategy=3-shot2025.02 | 39.3 | 40.2 | |
| Self-CoTModel=Qwen2.5-SLM2025.02 | 39.2 | 39.7 | |
| Auto-CoTModel=Qwen2.5-SLM, Prompting Strategy=3-shot2025.02 | 39.1 | 39.4 | |
| Naive w/ ChoiceModel=Llama3.1-SLM2025.02 | 38.7 | 38.8 | |
| Self-CoTModel=Llama3.1-SLM2025.02 | 35.4 | 36.2 | |
| PS-CoTModel=Llama3.1-SLM2025.02 | 29.1 | 30 | |
| Self-ConsistencyModel=Qwen2.5-SLM2025.02 | 28.8 | 29.3 | |
| Self-ConsistencyModel=Llama3.1-SLM2025.02 | 26.3 | 26.9 | |
| NaiveModel=Llama3.1-SLM2025.02 | 24 | 33.3 |