Can Small LLMs Learn a Robust Theory of Mind via RLVR? Investigating Generalization through the False-Belief Task

摘要

Recent advancements in large language models (LLMs) have demonstrated emergent capabilities in complex reasoning, largely spurred by rule-based Reinforcement Learning (RL) techniques applied during post-training. This has raised the question of whether similar methods can instill more nuanced, human-like social intelligence, such as a Theory of Mind (ToM), in LLMs. This paper investigates whether small- scale LLMs can acquire a robust and generalizable ToM capability through RL with verifiable rewards (RLVR). We conduct a systematic evaluation by training models on various combinations of prominent ToM benchmarks (HiToM, ExploreToM, FANToM) and testing for generalization on held-out benchmarks (e.g., Open- ToM). Our findings indicate that small LLMs struggle to develop a generic ToM capability. While performance on in-distribution tasks improves, this capability fails to transfer to unseen ToM tasks with different characteristics. Even observed out-of-distribution (OOD) performance improvements occur unpredictably across the training run, and don’t generalize across other OOD benchmarks. Furthermore, we conduct analysis to show that the learned behavior is likely a form of narrow overfitting rather than the acquisition of a true, abstract ToM capability.