METER: Evaluating Multi-Level Contextual Causal Reasoning in Large Language Models
For researchers studying LLM reasoning, this work provides a systematic benchmark and mechanistic analysis of causal reasoning limitations.
The paper introduces METER, a benchmark for evaluating LLMs across all three levels of causal reasoning under a unified context. Results show a significant performance decline as tasks ascend the causal hierarchy, with two failure modes identified: distraction by irrelevant factual information and reduced faithfulness to context.
Contextual causal reasoning is a critical yet challenging capability for Large Language Models (LLMs). Existing benchmarks, however, often evaluate this skill in fragmented settings, failing to ensure context consistency or cover the full causal hierarchy. To address this, we pioneer METER to systematically benchmark LLMs across all three levels of the causal ladder under a unified context setting. Our extensive evaluation of various LLMs reveals a significant decline in proficiency as tasks ascend the causal hierarchy. To diagnose this degradation, we conduct a deep mechanistic analysis via both error pattern identification and internal information flow tracing. Our analysis reveals two primary failure modes: (1) LLMs are susceptible to distraction by causally irrelevant but factually correct information at lower level of causality; and (2) as tasks ascend the causal hierarchy, faithfulness to the provided context degrades, leading to a reduced performance. We belive our work advances our understanding of the mechanisms behind LLM contextual causal reasoning and establishes a critical foundation for future research. Our code and dataset are available at https://github.com/SCUNLP/METER .