Frontier LLMs achieve 95-100% accuracy on AMC/AIME problems but recover far fewer distinct valid strategies than human references, while collectively generating 50 novel strategies.
hub Mixed citations
O lympiad B ench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems
Mixed citation behavior. Most common role is background (38%).
31
Pith papers citing it
Background
38% of classified citations
hub tools
citation-role summary
dataset 4
background 2
method 2
citation-polarity summary
representative citing papers
OptiVerse is a new benchmark spanning neglected optimization domains that shows LLMs suffer sharp accuracy drops on hard problems due to modeling and logic errors, with a Dual-View Auditor Agent proposed to improve performance.
Frontier LLMs struggle to discriminate data uncertainty from model uncertainty even when accurate, but a new benchmark and lightweight RL strategy improve attribution without sacrificing answer accuracy.
Evo-L2S uses multi-objective evolutionary model merging to produce reasoning models that cut generated chain-of-thought length by over 50% while preserving or improving accuracy on math benchmarks.
Prefix-RFT blends SFT and RFT via prefix sampling from demonstrations to outperform standalone SFT, RFT, and mixed-policy baselines on math reasoning problems.
Dynamic Gradient Gating monitors lm_head gradient norms to safely reuse rollout batches in RLVR, achieving up to 2.93x sample efficiency and 2.14x wall-clock speedup across math, ALFWorld, WebShop, and QA tasks.
PUMA detects reasoning-level semantic redundancy to enable early exit in chains of thought, achieving 26.2% average token reduction across five LRMs and five benchmarks while preserving accuracy and CoT quality.
PopuLoRA shows that co-evolving populations of LoRA adapters through cross-evaluated self-play can outperform compute-matched single-agent baselines on multiple code and math reasoning benchmarks.
GRLO shows RLHF from scratch on 5K open-ended prompts raises average performance from 24.1 to 63.1 across domains on Qwen3-4B-Base using 46x less data and 68x less compute than in-domain RLVR while remaining competitive with heavily post-trained models.
ICRL uses joint RL training of solver and critic with distribution-calibration re-weighting and role-wise advantage estimation to internalize critique into unassisted LLM performance, yielding 6.4-point gains on agentic tasks and 7.0 on math reasoning with Qwen3 models.
HölderPO unifies token-level aggregation in GRPO via the Hölder mean with a tunable p parameter and annealing schedule, delivering 54.9% average accuracy on math benchmarks and 93.8% success on ALFWorld.
EvoTD applies crossover for skill composition and parametric mutation for complexity scaling, filtered by a Zone of Proximal Development, to generate tasks that improve LLM reasoning generalization across models.
SORT turns all-wrong prompts into selective learning signals by weighting tokens more predictable under plan guidance from reference solutions, improving over GRPO on reasoning benchmarks especially for weaker models.
ExpThink reduces average CoT response length by up to 77% while improving accuracy on math benchmarks via experience-guided reward shaping and difficulty-adaptive advantage in RL.
RLCM trains LLMs with a margin-enhanced process reward that widens the gap between correct and incorrect reasoning steps, improving calibration on math, code, logic, and science tasks without hurting accuracy.
PDDL planning problems are used to generate about one million precise reasoning steps for training Process Reward Models, and adding this data to existing datasets improves LLM performance on both mathematical and non-mathematical reasoning benchmarks.
MEDS improves LLM RL performance by up to 4.13 pass@1 and 4.37 pass@128 points by dynamically penalizing rollouts matching prevalent historical error clusters identified via memory-stored representations and density clustering.
NPR trains LLMs to reason in parallel via self-distilled RL, delivering up to 24.5% performance gains and 4.6x speedups with 100% genuine parallel execution on reasoning benchmarks.
Entropy Ratio Clipping introduces a global entropy-ratio constraint that stabilizes RL policy updates in LLM post-training beyond local PPO clipping.
Archer introduces response-level entropy normalization and differentiated clipping/KL regularization in RLVR to encourage exploration on reasoning tokens while stabilizing knowledge tokens, yielding gains in pass@1 and pass@K on reasoning benchmarks.
Entropy minimization on self-generated outputs elicits strong reasoning in pretrained LLMs, matching or exceeding supervised RL methods on benchmarks.
PRIME enables online process reward model updates in LLM RL using implicit rewards from rollouts and outcome labels, yielding 15.1% average gains on reasoning benchmarks and surpassing a stronger instruct model with 10% of the data.
LLMs fail at extended counting of repeated characters due to finite internal states, with abrupt errors persisting across model scales and inference methods.
MathArena is broadened into a maintained platform with new benchmarks for proofs, research questions, and formal verification, where GPT-5.5 scores 98% on 2026 USAMO and 74% on research-level tasks.
citing papers explorer
-
Beyond Accuracy: Evaluating Strategy Diversity in LLM Mathematical Reasoning
Frontier LLMs achieve 95-100% accuracy on AMC/AIME problems but recover far fewer distinct valid strategies than human references, while collectively generating 50 novel strategies.
-
OptiVerse: A Comprehensive Benchmark towards Optimization Problem Solving
OptiVerse is a new benchmark spanning neglected optimization domains that shows LLMs suffer sharp accuracy drops on hard problems due to modeling and logic errors, with a Dual-View Auditor Agent proposed to improve performance.
-
Beyond "I Don't Know": Evaluating LLM Self-Awareness in Discriminating Data and Model Uncertainty
Frontier LLMs struggle to discriminate data uncertainty from model uncertainty even when accurate, but a new benchmark and lightweight RL strategy improve attribution without sacrificing answer accuracy.
-
Multi-objective Evolutionary Merging Enables Efficient Reasoning Models
Evo-L2S uses multi-objective evolutionary model merging to produce reasoning models that cut generated chain-of-thought length by over 50% while preserving or improving accuracy on math benchmarks.
-
Blending Supervised and Reinforcement Fine-Tuning with Prefix Sampling
Prefix-RFT blends SFT and RFT via prefix sampling from demonstrations to outperform standalone SFT, RFT, and mixed-policy baselines on math reasoning problems.
-
When to Stop Reusing: Dynamic Gradient Gating for Sample-Efficient RLVR
Dynamic Gradient Gating monitors lm_head gradient norms to safely reuse rollout batches in RLVR, achieving up to 2.93x sample efficiency and 2.14x wall-clock speedup across math, ALFWorld, WebShop, and QA tasks.
-
Stop When Reasoning Converges: Semantic-Preserving Early Exit for Reasoning Models
PUMA detects reasoning-level semantic redundancy to enable early exit in chains of thought, achieving 26.2% average token reduction across five LRMs and five benchmarks while preserving accuracy and CoT quality.
-
PopuLoRA: Co-Evolving LLM Populations for Reasoning Self-Play
PopuLoRA shows that co-evolving populations of LoRA adapters through cross-evaluated self-play can outperform compute-matched single-agent baselines on multiple code and math reasoning benchmarks.
-
GRLO: Towards Generalizable Reinforcement Learning in Open-Ended Environments from Zero
GRLO shows RLHF from scratch on 5K open-ended prompts raises average performance from 24.1 to 63.1 across domains on Qwen3-4B-Base using 46x less data and 68x less compute than in-domain RLVR while remaining competitive with heavily post-trained models.
-
ICRL: Learning to Internalize Self-Critique with Reinforcement Learning
ICRL uses joint RL training of solver and critic with distribution-calibration re-weighting and role-wise advantage estimation to internalize critique into unassisted LLM performance, yielding 6.4-point gains on agentic tasks and 7.0 on math reasoning with Qwen3 models.
-
Holder Policy Optimisation
HölderPO unifies token-level aggregation in GRPO via the Hölder mean with a tunable p parameter and annealing schedule, delivering 54.9% average accuracy on math benchmarks and 93.8% success on ALFWorld.
-
Evolutionary Task Discovery: Advancing Reasoning Frontiers via Skill Composition and Complexity Scaling
EvoTD applies crossover for skill composition and parametric mutation for complexity scaling, filtered by a Zone of Proximal Development, to generate tasks that improve LLM reasoning generalization across models.
-
Selective Off-Policy Reference Tuning with Plan Guidance
SORT turns all-wrong prompts into selective learning signals by weighting tokens more predictable under plan guidance from reference solutions, improving over GRPO on reasoning benchmarks especially for weaker models.
-
ExpThink: Experience-Guided Reinforcement Learning for Adaptive Chain-of-Thought Compression
ExpThink reduces average CoT response length by up to 77% while improving accuracy on math benchmarks via experience-guided reward shaping and difficulty-adaptive advantage in RL.
-
Process Supervision of Confidence Margin for Calibrated LLM Reasoning
RLCM trains LLMs with a margin-enhanced process reward that widens the gap between correct and incorrect reasoning steps, improving calibration on math, code, logic, and science tasks without hurting accuracy.
-
Process Reward Models Meet Planning: Generating Precise and Scalable Datasets for Step-Level Rewards
PDDL planning problems are used to generate about one million precise reasoning steps for training Process Reward Models, and adding this data to existing datasets improves LLM performance on both mathematical and non-mathematical reasoning benchmarks.
-
The Past Is Not Past: Memory-Enhanced Dynamic Reward Shaping
MEDS improves LLM RL performance by up to 4.13 pass@1 and 4.37 pass@128 points by dynamically penalizing rollouts matching prevalent historical error clusters identified via memory-stored representations and density clustering.
-
Native Parallel Reasoner: Reasoning in Parallelism via Self-Distilled Reinforcement Learning
NPR trains LLMs to reason in parallel via self-distilled RL, delivering up to 24.5% performance gains and 4.6x speedups with 100% genuine parallel execution on reasoning benchmarks.
-
Entropy Ratio Clipping as a Soft Global Constraint for Stable Reinforcement Learning
Entropy Ratio Clipping introduces a global entropy-ratio constraint that stabilizes RL policy updates in LLM post-training beyond local PPO clipping.
-
Stabilizing Knowledge, Promoting Reasoning: Dual-Token Constraints for RLVR
Archer introduces response-level entropy normalization and differentiated clipping/KL regularization in RLVR to encourage exploration on reasoning tokens while stabilizing knowledge tokens, yielding gains in pass@1 and pass@K on reasoning benchmarks.
-
The Unreasonable Effectiveness of Entropy Minimization in LLM Reasoning
Entropy minimization on self-generated outputs elicits strong reasoning in pretrained LLMs, matching or exceeding supervised RL methods on benchmarks.
-
Process Reinforcement through Implicit Rewards
PRIME enables online process reward model updates in LLM RL using implicit rewards from rollouts and outcome labels, yielding 15.1% average gains on reasoning benchmarks and surpassing a stronger instruct model with 10% of the data.
-
Language models fail at extended rule following
LLMs fail at extended counting of repeated characters due to finite internal states, with abrupt errors persisting across model scales and inference methods.
-
Beyond Benchmarks: MathArena as an Evaluation Platform for Mathematics with LLMs
MathArena is broadened into a maintained platform with new benchmarks for proofs, research questions, and formal verification, where GPT-5.5 scores 98% on 2026 USAMO and 74% on research-level tasks.
-
Efficient Test-Time Scaling via Temporal Reasoning Aggregation
TRACE aggregates answer consistency and confidence trajectory over multiple reasoning steps to decide when to halt inference, reducing token usage by 25-30% while keeping accuracy within 1-2% of full reasoning.
-
Depth-Breadth Synergy in RLVR: Unlocking LLM Reasoning Gains with Adaptive Exploration
DARS adaptively increases rollouts on hard problems in RLVR to improve Pass@K, and when paired with batch scaling for breadth, achieves gains in both Pass@K and Pass@1 by treating depth and breadth as complementary exploration dimensions.
-
Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models
The paper unifies perspectives on Long CoT in reasoning LLMs by introducing a taxonomy, detailing characteristics of deep reasoning and reflection, and discussing emergence phenomena and future directions.
-
Mathematical Reasoning in Large Language Models: Benchmarks, Architectures, Evaluation, and Open Challenges
A literature survey synthesizing benchmarks, architectures, training strategies, and evaluation methods for mathematical reasoning in LLMs, based on roughly 120 papers.
- Dual-Cluster Memory Agent: Resolving Multi-Paradigm Ambiguity in Optimization Problem Solving
- OGER: A Robust Offline-Guided Exploration Reward for Hybrid Reinforcement Learning
- FeynmanBench: Benchmarking Multimodal LLMs on Diagrammatic Physics Reasoning