arXiv · cs.AI · Shanglin Yang, Zhan Shi · 11 小时前模型评分51
预注册研究：筛选大语言模型生成的假设可提升超出模型评审预期的真实性
Selecting LLM-Generated Hypotheses Raises Truth Beyond What an LLM Panel Expects: A Pre-registered Study on NAS-Bench-201 and HPOBench
导读
一项预注册研究在 NAS-Bench-201 和 HPOBench 两个完整表格化基准上检验大语言模型生成的假设，发现先起草大量条件交互断言、再在单独一次调用中筛选子集的两步做法，能提升超出模型评审预期的真实性。
分类置信度
中 · 待人工复核
- 只有单一媒体来源，需要人工复核。
来源级别：媒体与个人
应用领域：科学研究
事件类型：研究发布
证据摘录
- “Across a series of pre-registered stages with Claude and GPT generators, our main positive result is a simple two-stage step found by a harness optimizer: draft a large pool of claims, then select a subset in a separate call.”
- “Because kept claims recur across runs, the evidence weakens when runs and claims are resampled jointly, and it is not significant against Sonnet's own prior.”
- “Optimizing the harness for an LLM judge raised judge scores while sharply cutting precision for every generator.”
来源：arXiv · cs.AI · arxiv.org