Approximating human responses on Pew Research Center ATP Survey Wave 99
0.081WDHuman
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Human2024.07 | 0.081 | 0.327 | 0.83 | |
| Anthology (DP)Model=Mixtral-8x22B, Persona Conditioning=Anthology (DP), Persona Matching=n/a2024.07 | 0.215 | 1.422 | 0.604 | |
| BioModel=Mixtral-8x22B, Persona Conditioning=Bio, Persona Matching=n/a2024.07 | 0.237 | 1.092 | 0.687 | |
| Anthology (NA)Model=Llama-3-70B, Persona Conditioning=Anthology (NA), Persona Matching=greedy2024.07 | 0.288 | 0.765 | 0.744 | |
| BioModel=Llama-3-70B, Persona Conditioning=Bio, Persona Matching=n/a2024.07 | 0.296 | 0.809 | 0.733 | |
| Anthology (DP)Model=Llama-3-70B, Persona Conditioning=Anthology (DP), Persona Matching=n/a2024.07 | 0.302 | 1.14 | 0.669 | |
| Anthology (NA)Model=Mixtral-8x22B, Persona Conditioning=Anthology (NA), Persona Matching=greedy2024.07 | 0.32 | 0.951 | 0.71 | |
| QAModel=Llama-3-70B, Persona Conditioning=QA, Persona Matching=n/a2024.07 | 0.327 | 0.767 | 0.74 | |
| Anthology (NA)Model=Llama-3-70B, Persona Conditioning=Anthology (NA), Persona Matching=max weight2024.07 | 0.327 | 0.686 | 0.756 | |
| Anthology (NA)Model=Mixtral-8x22B, Persona Conditioning=Anthology (NA), Persona Matching=max weight2024.07 | 0.353 | 0.843 | 0.729 | |
| QAModel=Mixtral-8x22B, Persona Conditioning=QA, Persona Matching=n/a2024.07 | 0.395 | 1.086 | 0.684 |