Prompt Repetition Experiments
收藏DataCite Commons2026-05-05 更新2026-05-07 收录
下载链接：
https://zenodo.org/doi/10.5281/zenodo.20042392
下载链接
链接失效反馈官方服务：
资源简介：
30 agent sessions (4,306 messages) across three pre-registered controlled experiments testing whether prompt repetition improves LLM-agent task completion on structured software-engineering tasks. Null results across all three experiments; identifies ceiling effects and rubric-runner co-design failure as distinct evaluation failure modes. Supplementary dataset for: Clouatre, H. (2026). Ceiling Effects and Convergence: Null Results for Instruction Repetition in LLM-Agent Pipelines. Preprint: https://doi.org/10.5281/zenodo.20039271
提供机构：
Zenodo
创建时间：
2026-05-05
