芝加哥大学论文：模型自述的关键决策因素与实际不符
芝加哥大学做了个挺反直觉的实验：问模型自己在乎什么，它答得还不如瞎猜，得靠对比多份回答才能找到真正的分歧点，方法也便宜。
芝加哥大学一篇论文测试了 LLM 能否预测自己回答中的关键分歧点（cruxes）。实验让模型多次预测 59 场 March Madness 比赛和 15 个 Metaculus 事件，再对比多份回答找出争议事实。结果显示，模型预测的关键点只与真实分歧簇共享 13.4%，低于随机打乱得到的 39.2%；单份回答也只覆盖 28.6% 的真实关键点，低于 37.8% 的基线。论文提出用 reranker 对比多份回答来定位分歧，计算量估计比本地 LLM judge 少 86 倍。
New University of Chicago paper suggests that an LLM's own account of what matters to its decision is a poor guide to what actually varies across its answers.
That LLMs fail to predict the key questions their own answers split on, which only comparing many answers reveals.
When an LLM recommends something, you want its cruxes, the few sub-questions that drove its conclusion. Asking the model upfront, or reading 1 of its answers, is cheap but may miss what really drives it.
They had models forecast 59 March Madness games and 15 Metaculus events many times, then compared pairs of forecasts to find disputed facts. A reranker made that matching cheap, at an estimated 86 times less compute than a local LLM judge.
In March Madness, predicted cruxes shared 13.4% of clusters with the real ones, below the 39.2% a random shuffle gives. A single answer did better but covered only 28.6% of real cruxes, against a 37.8% baseline.
For high-stakes decisions, sample many answers and look where they disagree, instead of trusting the model's own list of what matters.
