Humans and LLMs Diverge on Probabilistic Inferences
This research highlights a critical gap in LLM reasoning capabilities, specifically their inability to mimic human-like probabilistic judgments, which is important for developing more human-aligned AI.
This paper investigates how Large Language Models (LLMs) perform on probabilistic inferences compared to humans. The authors created ProbCOPA, a dataset of 210 probabilistic inferences, and found that LLMs consistently fail to produce human-like distributions of responses, indicating a divergence in their probabilistic reasoning.
Human reasoning often involves working over limited information to arrive at probabilistic conclusions. In its simplest form, this involves making an inference that is not strictly entailed by a premise, but rather only likely given the premise. While reasoning LLMs have demonstrated strong performance on logical and mathematical tasks, their behavior on such open-ended, non-deterministic inferences remains largely unexplored. We introduce ProbCOPA, a dataset of 210 handcrafted probabilistic inferences in English, each annotated for inference likelihood by 25--30 human participants. We find that human responses are graded and varied, revealing probabilistic judgments of the inferences in our dataset. Comparing these judgments with responses from eight state-of-the-art reasoning LLMs, we show that models consistently fail to produce human-like distributions. Finally, analyzing LLM reasoning chains, we find evidence of a common reasoning pattern used to evaluate such inferences. Our findings reveal persistent differences between humans and LLMs, and underscore the need to evaluate reasoning beyond deterministic settings.