Bayesian Elicitation with LLMs: Model Size Helps, Extra "Reasoning" Doesn't Always
This addresses the problem of unreliable uncertainty estimates from LLMs for decision-makers, showing they require statistical correction, but it is incremental as it builds on existing elicitation and calibration methods.
The study tested whether large language models (LLMs) can accurately estimate unknown quantities with uncertainty, finding that larger models improved accuracy but extra reasoning effort did not, and all models were overconfident with 95% intervals containing the true value only 9-44% of the time, though recalibration techniques could correct this.
Large language models (LLMs) have been proposed as alternatives to human experts for estimating unknown quantities with associated uncertainty, a process known as Bayesian elicitation. We test this by asking eleven LLMs to estimate population statistics, such as health prevalence rates, personality trait distributions, and labor market figures, and to express their uncertainty as 95\% credible intervals. We vary each model's reasoning effort (low, medium, high) to test whether more "thinking" improves results. Our findings reveal three key results. First, larger, more capable models produce more accurate estimates, but increasing reasoning effort provides no consistent benefit. Second, all models are severely overconfident: their 95\% intervals contain the true value only 9--44\% of the time, far below the expected 95\%. Third, a statistical recalibration technique called conformal prediction can correct this overconfidence, expanding the intervals to achieve the intended coverage. In a preliminary experiment, giving models web search access degraded predictions for already-accurate models, while modestly improving predictions for weaker ones. Models performed well on commonly discussed topics but struggled with specialized health data. These results indicate that LLM uncertainty estimates require statistical correction before they can be used in decision-making.