SOTAVerified

Mathematical Reasoning

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 51–100 of 805 papers

TitleStatusHype
Spurious Rewards: Rethinking Training Signals in RLVRCode3
MathArena: Evaluating LLMs on Uncontaminated Math CompetitionsCode3
Step-DPO: Step-wise Preference Optimization for Long-chain Reasoning of LLMsCode3
Self-Refine: Iterative Refinement with Self-FeedbackCode3
AlphaMath Almost Zero: Process Supervision without ProcessCode3
Self-rewarding correction for mathematical reasoningCode3
Reinforcement Learning for Reasoning in Large Language Models with One Training ExampleCode3
Reinforcement Learning for Reasoning in Small LLMs: What Works and What Doesn'tCode3
MuMath-Code: Combining Tool-Use Large Language Models with Multi-perspective Data Augmentation for Mathematical ReasoningCode3
SkyMath: Technical ReportCode3
MuggleMath: Assessing the Impact of Query and Response Augmentation on Math ReasoningCode2
DRLE: Decentralized Reinforcement Learning at the Edge for Traffic Light Control in the IoVCode2
A Survey of Deep Learning for Mathematical ReasoningCode2
Preference Optimization for Reasoning with Pseudo FeedbackCode2
Initialization using Update Approximation is a Silver Bullet for Extremely Efficient Low-Rank Fine-TuningCode2
Physics of Language Models: Part 2.1, Grade-School Math and the Hidden Reasoning ProcessCode2
ProcessBench: Identifying Process Errors in Mathematical ReasoningCode2
Omni-MATH: A Universal Olympiad Level Mathematic Benchmark For Large Language ModelsCode2
DisCO: Reinforcing Large Reasoning Models with Discriminative Constrained OptimizationCode2
Agent RL Scaling Law: Agent RL with Spontaneous Code Execution for Mathematical Problem SolvingCode2
Optimizing Anytime Reasoning via Budget Relative Policy OptimizationCode2
SOLO: A Single Transformer for Scalable Vision-Language ModelingCode2
Intelligent Go-Explore: Standing on the Shoulders of Giant Foundation ModelsCode2
GSM-Plus: A Comprehensive Benchmark for Evaluating the Robustness of LLMs as Mathematical Problem SolversCode2
O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning PruningCode2
Offline Reinforcement Learning for LLM Multi-Step ReasoningCode2
Reformatted AlignmentCode2
MINT-CoT: Enabling Interleaved Visual Tokens in Mathematical Chain-of-Thought ReasoningCode2
Gödel Agent: A Self-Referential Agent Framework for Recursive Self-ImprovementCode2
MetaMath: Bootstrap Your Own Mathematical Questions for Large Language ModelsCode2
Measuring Multimodal Mathematical Reasoning with MATH-Vision DatasetCode2
Confucius3-Math: A Lightweight High-Performance Reasoning LLM for Chinese K-12 Mathematics LearningCode2
Critique Fine-Tuning: Learning to Critique is More Effective than Learning to ImitateCode2
MegaMath: Pushing the Limits of Open Math CorporaCode2
FormalMATH: Benchmarking Formal Mathematical Reasoning of Large Language ModelsCode2
MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual ContextsCode2
Flow of Reasoning:Training LLMs for Divergent Problem Solving with Minimal ExamplesCode2
Compression Represents Intelligence LinearlyCode2
MathCoder: Seamless Code Integration in LLMs for Enhanced Mathematical ReasoningCode2
CoRT: Code-integrated Reasoning within ThinkingCode2
Math-LLaVA: Bootstrapping Mathematical Reasoning for Multimodal Large Language ModelsCode2
CRAFT: Customizing LLMs by Creating and Retrieving from Specialized ToolsetsCode2
MathPile: A Billion-Token-Scale Pretraining Corpus for MathCode2
An Expression Tree Decoding Strategy for Mathematical Equation GenerationCode2
MathOdyssey: Benchmarking Mathematical Problem-Solving Skills in Large Language Models Using Odyssey Math DataCode2
Measuring Mathematical Problem Solving With the MATH DatasetCode2
Multi-View Reasoning: Consistent Contrastive Learning for Math Word ProblemCode2
Exploring the Limit of Outcome Reward for Learning Mathematical ReasoningCode2
Benchmarking Benchmark Leakage in Large Language ModelsCode2
MAmmoTH: Building Math Generalist Models through Hybrid Instruction TuningCode2
Show:102550
← PrevPage 2 of 17Next →
All datasetsAIME24FrontierMathLila (IID)Lila (OOD)PGPS9KAMC23GeoQAMath500UniGeoUniGeo (PRV)

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1XolverAcc94.4—Unverified
2DeepSeek-r1Acc79.8—Unverified
3Openai-o1Acc74.4—Unverified
4Openai-o1-miniAcc70—Unverified
5Search-o1Acc56.7—Unverified
6s1-32BAcc56.7—Unverified
7Openai-o1-previewAcc44.6—Unverified
8Qwen2.5-72B-InstructAcc23.3—Unverified
9Claude3.5-SonnetAcc16—Unverified
#ModelMetricClaimedVerifiedStatus
1o3Accuracy0.25—Unverified
2Gemini 1.5 Pro (002)Accuracy0.02—Unverified
3GPT-4oAccuracy0.01—Unverified
4o1-miniAccuracy0.01—Unverified
5o1-previewAccuracy0.01—Unverified
6Claude 3.5 SonnetAccuracy0.01—Unverified
#ModelMetricClaimedVerifiedStatus
1Codex (Few-Shot, 175B)Accuracy0.6—Unverified
2Bhāskara-P (Fine-tuned, 2.7B)Accuracy0.48—Unverified
3Neo-P (Fine-tuned, 2.7B)Accuracy0.39—Unverified
4GPT-3 (Few-Shot, 175B)Accuracy0.38—Unverified
5Bhāskara-A (Fine-tuned, 2.7B)Accuracy0.25—Unverified
6Neo-A (Fine-tuned, 2.7B)Accuracy0.2—Unverified
#ModelMetricClaimedVerifiedStatus
1Codex (Few-Shot, 175B)Accuracy0.59—Unverified
2Bhāskara-P (Fine-tuned, 2.7B)Accuracy0.45—Unverified
3GPT-3 (Few-Shot, 175B)Accuracy0.38—Unverified
4Bhāskara-A (Fine-tuned, 2.7B)Accuracy0.27—Unverified
5Neo-P (Fine-tuned, 2.7B)Accuracy0.24—Unverified
6Neo-A (Fine-tuned, 2.7B)Accuracy0.18—Unverified
#ModelMetricClaimedVerifiedStatus
1GOLDCompletion accuracy65.8—Unverified
2PGPSNetCompletion accuracy62.7—Unverified
3GAPSCompletion accuracy61.2—Unverified
4Inter-GPSCompletion accuracy59.8—Unverified
5GeoformerCompletion accuracy35.6—Unverified
6NGSCompletion accuracy34.1—Unverified
#ModelMetricClaimedVerifiedStatus
1QWQ-32B-previewAcc82.5—Unverified
2Math-MasterAcc82—Unverified
3Qwen2.5-Math-7B-instructAcc62.5—Unverified
#ModelMetricClaimedVerifiedStatus
1GOLDAccuracy (%)75.2—Unverified
2GAPSAccuracy (%)67.8—Unverified
#ModelMetricClaimedVerifiedStatus
1Search-o1Acc86.4—Unverified
#ModelMetricClaimedVerifiedStatus
1GOLDAccuracy (%)98.5—Unverified
#ModelMetricClaimedVerifiedStatus
1GAPSAccuracy (%)97.5—Unverified