LingBench++: A Linguistically-Informed Benchmark and Reasoning Framework for Multi-Step and Cross-Cultural Inference with LLMs
This work addresses the need for linguistically grounded and culturally informed reasoning benchmarks for LLMs, though it is incremental as it builds on prior benchmarks by adding structured evaluation and multi-agent methods.
The authors tackled the problem of evaluating large language models on complex linguistic tasks by introducing LingBench++, a benchmark with structured reasoning traces and stepwise evaluation across over 90 low-resource and cross-cultural languages, and demonstrated that multi-agent models with external knowledge and iterative reasoning outperform single-pass approaches in accuracy and interpretability.
We propose LingBench++, a linguistically-informed benchmark and reasoning framework designed to evaluate large language models (LLMs) on complex linguistic tasks inspired by the International Linguistics Olympiad (IOL). Unlike prior benchmarks that focus solely on final answer accuracy, LingBench++ provides structured reasoning traces, stepwise evaluation protocols, and rich typological metadata across over 90 low-resource and cross-cultural languages. We further develop a multi-agent architecture integrating grammatical knowledge retrieval, tool-augmented reasoning, and deliberate hypothesis testing. Through systematic comparisons of baseline and our proposed agentic models, we demonstrate that models equipped with external knowledge sources and iterative reasoning outperform single-pass approaches in both accuracy and interpretability. LingBench++ offers a comprehensive foundation for advancing linguistically grounded, culturally informed, and cognitively plausible reasoning in LLMs.