Results: Deliberative Synthesis vs. Single-Model and Majority-Vote Baselines on GPQA Diamond
收藏DataCite Commons2026-05-03 更新2026-05-07 收录
下载链接:
https://zenodo.org/doi/10.5281/zenodo.19992033
下载链接
链接失效反馈官方服务:
资源简介:
Experimental results for preregistered study (DOI 10.5281/zenodo.19991489). N=198 GPQA Diamond questions. H1 supported: synthesis > single model (p=0.000072). H3 supported: vote > single model (p=0.000062). H2 not supported: synthesis does not significantly outperform vote (p=0.42). Total cost $0.13.
提供机构:
Zenodo
创建时间:
2026-05-03



