约翰霍普金斯大学论文：LLM训练数据中单例事实促进科学发现
Strange findings here.
这篇论文揭示了单例数据对模型创造力的关键作用，为科学 AI 的数据构建提供了反直觉但可验证的指导原则，值得研究者关注。
Strange findings here.
此处发现异常。
New Johns Hopkins paper finds that LLMs make more scientific discoveries when more of their training facts appear only once.
约翰霍普金斯大学的新论文发现，当训练事实中更多仅出现一次时，大语言模型（LLMs）能做出更多的科学发现。
discovery rates rose steadily as the share of once-seen facts in training went from 0% to 100%, for 2 model families.
对于两种模型系列，随着训练中仅见过一次的事实占比从 0% 增加到 100%，发现率稳步上升。
Discovery and hallucination both mean producing something the model never saw in training. Earlier theory tied hallucination to "monofacts", facts seen exactly once, and treated that unseen space as mostly false.
“发现”和“幻觉”都指模型生成了其在训练中从未见过的内容。早期的理论将幻觉与“单例事实”（monofacts，即恰好只出现过一次的事实）联系起来，并将这种未见过的空间主要视为虚假内容。
The authors extend that theory to show monofacts also set how much a model can discover. They fine-tuned gpt-oss-20b and gemma-3-12b-it on materials and DNA datasets, varying the monofact share from 0% to 100%.
作者扩展了该理论，表明单例事实也决定了模型能够进行多少发现。他们在材料和 DNA 数据集上微调了 gpt-oss-20b 和 gemma-3-12b-it，将单例事实的占比从 0% 变化到 100%。
Discovery, scored as exactly recovering materials or DNA sequences held out from training, rose steadily with the monofact share in every setting. Training for 3 or 5 epochs, which turns single sightings into repeats, made the trend disappear.
在每种设置下，通过精确恢复从训练中排除的材料或 DNA 序列来衡量的“发现”，随单例事实占比的增加而稳步上升。进行 3 或 5 个轮次（epochs）的训练会将单次目击变为重复，从而使这一趋势消失。
When building science models, keep rare one-off examples and avoid repeating data, and expect more hallucinations to come with more discoveries.
在构建科学模型时，应保留稀有的单次示例并避免重复数据，同时要预期更多的幻觉会伴随更多的发现而来。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力