Rohan Paul (@rohanpaul_ai) · 2026/10/11 19:23:08
亚马逊提出 LLM 作为科研世界模型：利用历史实验记录预测训练收益，将排序相关性提升至 0.774
亚马逊最新研究提出将大语言模型（LLM）用作“科研世界模型”，旨在通过预测实验结果来优化 GPU 资源分配。该方法利用同一配置下的历史实验记录进行上下文学习，在 9 种设置、2653 条真实记录上验证，使预测与实际结果的平均排名相关性从 0.506 显著提升至 0.774。研究表明，为 AI 研究智能体提供丰富的失败与成功日志，比单纯增加推理算力更能有效指导后续实验选择。
报道全文原始报道全文
Rohan Paul (@rohanpaul_ai) · 2026-10-11 11
New Amazon paper shows that an LLM can act as a research world model, predicting whether a training change will help before anyone spends GPU time on it.
AI research agents can propose experiments far faster than teams can afford to run them. Choosing what gets GPU time means guessing outcomes in advance.
They used an LLM as a research world model that predicts an experiment's gain before it runs. They tested it on 2,653 real experiment records from 9 setups, from pretraining to inference.
Past records from the same setup raised average ranking correlation with actual results from 0.506 to 0.774 across 5 setups. On the OLMo3-100M setup, adding records at low reasoning effort scored 0.892, while max effort without records reached 0.648.
If you run research agents, log every experiment, including failures, and feed those records to whatever model picks the next run.
