LGMT: Logic-Grounded Metamorphic Testing for Evaluating the Reasoning Reliability of LLMs

In three linesLGMT is an oracle-free evaluation framework using first-order logic to test LLM reasoning reliability. By deriving metamorphic relations from formal logical equivalences, it constructs semantically invariant test cases. Experiments on 6 state-of-the-art LLMs expose hidden defects missed by traditional static benchmarks.

Summary generated by Claude — human-verified