arXiv:2605. 19723v2 Announce Type: replace-cross Abstract: Mathematical reasoning is essential for problem-solving in education, science, and industry, serving as a crucial benchmark for evaluating artificial intelligence systems.
arXiv:2501. 11790v5 Announce Type: replace-cross Abstract: Recent studies have raised significant concerns regarding the reliability of current mathematics benchmarks, highlighting issues such as simplistic design and potential data contamination.
arXiv:2607. 05992v1 Announce Type: cross Abstract: Mathematical reasoning has become a central task for evaluating and tuning reasoning Large Language Models (LLMs), yet existing benchmarks remain heavily biased toward high-resource languages, with English and Chinese dominating both pre-training corpora and evaluation suites.
AdaR is a framework designed to enhance large language models (LLMs) with adaptive reasoning for mathematical tasks. It identifies and mitigates spurious reasoning—where models rely on superficial correlations—by generating logically equivalent queries and training with Reinforcement Learning with Verifiable Rewards (RLVR) to penalize incorrect logic and promote adaptive logic. The approach includes extracting problem‑solving logic, executing code to verify answers, and applying sanity checks, resulting in significant gains in mathematical reasoning performance and improved data efficiency.
The paper introduces an epistemically and formally grounded ensemble (EFG) of large language model judges to evaluate autoformalization tasks in formal mathematics. It defines four criteria—logical preservation, mathematical consistency, formal quality, and formal validity—to provide a transparent, multi‑granular assessment. Experiments show that this ensemble outperforms coarse‑grained models, offering a scalable and interpretable proxy for evaluating formal mathematical reasoning.
arXiv:2606. 26797v1 Announce Type: new Abstract: Supervised fine-tuning (SFT) on a small, high-quality set of long reasoning traces is an effective approach for eliciting strong reasoning capabilities in Large Language Models (LLMs).
The paper investigates whether reasoning representations—explanations for large language model outputs—aid humans in evaluating those outputs. A controlled human study tested six reasoning formats across tasks of varying complexity, measuring structural understanding, error detection, and trust calibration. Results revealed a mismatch: participants favored planning- and decomposition-based representations, yet simpler chain-of-thought traces better supported verification, trust, and interpretability, while preferred formats increased calibration risks.
Supervised fine-tuning (SFT) on a small, high-quality set of long reasoning traces is an effective approach for eliciting strong reasoning capabilities in Large Language Models (LLMs). However, existing methods for curating high-quality SFT data rely heavily on strong reasoning models to filter examples based on diversity and difficulty, making the curation process costly while often yielding suboptimal data quality.
The paper introduces CodeRQ-Bench, the first benchmark for assessing large language model reasoning quality across coding tasks such as generation, summarization, and classification. It analyzes over a thousand mismatches from existing evaluators, identifies recurring limitations, and derives design insights that lead to a new two‑stage evaluator, VERA. Experiments show VERA outperforms strong baselines, improving AUCROC by up to 0.26 and AUPRC by up to 0.21 on four datasets.
The paper argues that large language models (LLMs) organize their internal mathematical reasoning by reusable reasoning approaches rather than by the benchmark topics they are tested on. Using a generation‑replay protocol, the authors extract activation‑importance signatures from eight models across five math sources, cluster these signatures, and find that the resulting groups align more closely with reasoning approaches than with topics. The study shows that changing the requested reasoning approach shifts cluster assignments, while paraphrasing the prompt does not, underscoring the primacy of approach over topic in LLM reasoning.
arXiv:2606. 07515v1 Announce Type: cross Abstract: We investigate the probabilistic reasoning capabilities of large language models through a controlled benchmarking study on discrete probability problems.
Diagrams are widely used to support logical reasoning, and prior studies suggest that representations such as Euler diagrams can improve human reasoning performance. Recent work has also explored their effects on large language models (LLMs).