Kyubyung Chae, Jewon Yeom, Jeongjae Park, Seunghyun Bae, Ijun Jang, Hyunbin Jin, Jinkwan Jang, Taesup Kim. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Improving the reasoning abilities of large language models (LLMs) has largely relied on iterative self-training with model-generated data.While effective at boosting accuracy, existing approaches primarily reinforce successful reasoning paths, incurring a substantial calibration cost: models become overconfident and lose the ability to represent uncertainty.This failure has been characterized as a form of model collapse in alignment, where predictive distributions degenerate toward low-variance point estimates.We address this issue by reframing open-ended reasoning training as an epistemic learning problem, in which models must learn not only how to reason, but also when their reasoning should be trusted.We propose epistemically-calibrated reasoning (EPICAR) as a training objective that jointly optimizes reasoning performance and calibration, and instantiate it within an iterative supervised fine-tuning framework using explicitly extracted meta-cognitive self-evaluation signals.Experiments on Llama-3 and Qwen-3 families demonstrate that our approach achieves Pareto-superiority over standard baselines in both accuracy and calibration, particularly in models with sufficient reasoning capacity (e.g., 3B+).This framework generalizes effectively to OOD mathematical reasoning (GSM8K) and code generation (MBPP).Ultimately, our approach enables a 3× reduction in the overall inference compute budget, matching the K = 30 majority-vote performance of STaR with only K = 10 confidence-weighted samples, entirely without the multi-model overhead of external verifiers.

Large language models (LLMs) trained on massive corpora demonstrate impressive capabilities in a wide range of tasks. While there are ongoing efforts to adapt these models to languages beyond English, the attention given to their evaluation methodologies remains limited. Current multilingual benchmarks often rely on back translations or re-implementations of English tests, limiting their capacity to capture unique cultural and linguistic nuances. To bridge this gap for the Korean language, we introduce the HAE-RAE Bench, a dataset curated to challenge models lacking Korean cultural and contextual depth. The dataset encompasses six downstream tasks across four domains: vocabulary, history, general knowledge, and reading comprehension. Unlike traditional evaluation suites focused on token and sequence classification or mathematical and logical reasoning, the HAE-RAE Bench emphasizes a model’s aptitude for recalling Korean-specific knowledge and cultural contexts. Comparative analysis with prior Korean benchmarks indicates that the HAE-RAE Bench presents a greater challenge to non-Korean models by disturbing abilities and knowledge learned from English being transferred.