Sangmook Lee, Minbeom Kim, Jeonghye Kim, Dohyung Kim, Sojeong Rhee, Kyomin Jung · Jun 29, 2026 · Citations: 0
Pairwise PreferenceMath
Using a human-calibrated LLM judge framework, we show that prior diversity measures are unreliable proxies for approach-level diversity, and this mismatch carries over to diversity-aware RLVR, where target metrics are preserved while…
However, optimizing an LLM judge diversity reward during training causes the policy to exploit judge-specific preferences rather than broaden its approaches, leaving direct optimization of approach-level diversity as an open problem.
Nikolay Georgiev, Maria Drencheva, Kseniia Ibragimova, Ivo Petrov, Dimitar I. Dimitrov, Martin Vechev · Jun 29, 2026 · Citations: 0
Pairwise PreferenceAutomatic MetricsMath
As agentic AI systems tackle more complex mathematical tasks, they increasingly rely on information retrieval (IR) to search problem databases, theorem libraries, and educational resources.
Importantly, we show that general-purpose IR benchmarks such as MTEB do not reliably predict mathematical performance, especially for recent embedding models, highlighting the need for math-specific retrieval benchmarks.
Jaeyong Ko, Pilsung Kang, Yukyung Lee · Jun 24, 2026 · Citations: 0
Pairwise PreferenceAutomatic MetricsMath
Across seven models and three mathematical reasoning benchmarks (GSM1K, MATH500, AIME 2025), cliff tokens act as failure triggers; deleting the first cliff token and resampling recovers pass@64 to 1.0, while keeping it limits recovery to…
Trained on GSM8K, Cliff-DPO improves accuracy across benchmarks by up to +6.6.
Yutong Yin, Mingyu Jin, Jin Pan, Changyi Yang, Zijie Xia, Dhruv Pai · Jun 24, 2026 · Citations: 0
Automatic MetricsMath
On mathematical reasoning benchmarks, LBR improves both Pass@1 and Pass@32 over discrete chain-of-thought, vanilla discrete-token RLVR, and RL-compatible soft-token branching baselines.
In a six-variant Qwen3 math reasoning benchmark with a uniform 200-step training budget for all trained variants, we use pass@8 as the primary problem-level solve-rate metric.
On Teacher-TopK/LSM, Block64 gives the best four-benchmark mean pass@8 among trained students.
Junyi Yao, Zihao Zheng, Baichuan Li · Jun 15, 2026 · Citations: 0
Rubric RatingAutomatic MetricsMath
We propose a lightweight diagnostic based on the gap between solving-oriented and pedagogy-oriented benchmark performance.
Using public MathTutorBench leaderboard results, we show that these dimensions are only partially aligned: across eight publicly reported models, the correlation between solving and pedagogy composites is 0.421, and several models shift…
Therefore, we present Seeded Elimination with Adaptive LLM-as-a-Meta-Judge, a self-improving evaluation protocol for extracting latent ranking signal from saturated benchmarks.
We evaluate SEAL on multiple saturated benchmarks covering code generation, mathematical reasoning, knowledge-intensive question answering, and tool-use agent task completion.
Jinyang Wu, Guocheng Zhai, Ruihan Jin, Yuhao Shen, Zhengxi Lu, Fan Zhang · May 21, 2026 · Citations: 0
Expert VerificationAutomatic MetricsMathCoding
In this paper, we present Maestro (Multimodal Agent for Expert-Skill Targeted Reinforced Orchestration), a Reinforcement Learning (RL)-driven orchestration framework that reframes heterogeneous multimodal tasks as a sequential…
We evaluate Maestro across ten representative multimodal benchmarks spanning mathematical reasoning, chart understanding, high-resolution perception, and domain-specific analysis.
Mingkai Deng, Jinyu Hou, Lara Sá Neves, Varad Pimpalkhute, Taylor W. Killian, Zhengzhong Liu · May 21, 2026 · Citations: 0
Automatic MetricsMath
To test this, we develop SR^2AM (Self-Regulated Simulative Reasoning Agentic LLM), realizing both as distinct stages within an LLM's chain-of-thought, with the LLM as world model.
Across math, science, tabular analysis, and web information seeking, v0.1-8B and v1.0-30B achieve Pass@1 competitive with 120-355B and 685B-1T parameter systems respectively, while v1.0-30B uses 25.8-95.3% fewer reasoning tokens than…
Xitai Jiang, Zihan Tang, Wenze Lin, Yang Yue, Shenzhi Wang, Gao Huang · May 21, 2026 · Citations: 0
Rubric RatingAutomatic MetricsMath
Across seven mathematical reasoning benchmarks, SCRL outperforms strong curriculum-learning baselines, improving average accuracy over GRPO by +4.1 points on Qwen3-4B-Base and +1.9 points on Qwen3-14B-Base.
Across four state-of-the-art reasoning models, the proposed method substantially amplifies output length, achieving up to a 26.1x increase on the MATH benchmark and consistently outperforming benign and manually crafted missing-premise…
Zheng Fang, Xiaosen Wang, Shenyi Zhang, Shaokang Wang, Zhijin Ge · May 6, 2026 · Citations: 0
Red TeamMath
These results demonstrate that dense waveform updates are largely redundant, and we advocate that future audio jailbreak and safety alignment research should further leverage this heterogeneous token-level gradient structure.
Tianyang Han, Hengyu Shi, Junjie Hu, Xu Yang, Zhiling Wang, Junhao Su · May 5, 2026 · Citations: 0
Rubric RatingAutomatic MetricsMathLaw
Extensive experiments on code and math benchmarks show that this executor-grounded reasoning reward improves the two-stage planner-executor system over execution-only training, suggesting that reasoning supervision should evaluate not only…
