Majority Voting

Follow this topic Ask about this topic

Momentum

10 papers in the last four weeks, up 25% on the four weeks before. 0.1% of all new papers.

Jul 6Week of Sep 21

Latest papers 105

All topics

Consolidating Rewarded Perturbations for LLM Post-Training
May 29, 2026Zheyu Zhang, Shuo Yang, Gjergji KasneciLarge Language Model Reinforcement Learning Post-Training
Technical University of Munich
Diversity Matters: Revisiting Test-Time Compute in Vision-Language Models
May 29, 2026Yijie Tong, Yifan Hou, Shaobo Cui +2Recent Vision-Language Models Test Time
ETH Zürich · Shanghai Jiao Tong University · EPFL
Nine Judges, Two Effective Votes: Correlated Errors Undermine LLM Evaluation Panels
May 28, 2026Guneet KohliLlm-As-A-Judge Judges
Apple
The incremental voter model: mean-field analysis and convergence to equilibrium
May 27, 2026Fei Cao, Xiaoqian GongMean-Field Limit Majority Voting
Amherst College - Department of Mathematics, Amherst, MA 01002, USA
Forecasting Japanese elections: A nonlinear machine-learning approach
May 26, 2026Sota Kato, Xuan Luo, Budrul Ahsan +2Elections Decision Trees
International University of Japan · The Tokyo Foundation · IBM Japan +2
Minimax Limits of k-Fold Cross-Validation via Majority
May 25, 2026Ido Nachum, Rüdiger Urbanke, Thomas WeinbergerCross Validation Empirical Risk Minimization
University of Haifa, Haifa, Israel. · EPFL, Lausanne, Switzerland.
ARBITER: Reasoning Trajectory Basins and Majority Vote Failures in Test-Time Sampling
May 25, 2026Meng Cai, Lars Kulik, Farhana ChoudhuryArbitration Majority Voting
School of Computing and Information Systems University of Melbourne
The Communication Complexity of Instant-Runoff Voting
May 22, 2026Élie de Panafieu, François Durand, Jérôme LangMajority Voting Maximum Independent Set
LAMSADE, CNRS · 1Nokia Bell Labs · 2CNRS, LAMSADE, Universit´e Paris-Dauphine, PSL, France
Efficient Elicitation of Collective Disagreements
May 19, 2026Mohamed Ouaguenouni, Felipe Garrido-Lucero, Umberto Grandi +2Disagreement Majority Voting
IRIT, Université Toulouse Capitole, Toulouse, France · Center for Collective Learning, IAST, Toulouse School of Economics, Toulouse, France · Center for Collective Learning, CIAS, Corvinus University of Budapest, Budapest, Hungary +2
Detecting and Mitigating the Correct-Answer Extinction Window in Test-Time Reinforcement Learning with Majority Voting
May 19, 2026Hongxiang Lin, Zhirui Kuai, Erpeng Xue +1Reward Functions Mathematical Reasoning Benchmarks
Meituan
Swimming with Whales: Analysis of Power Imbalances in Stake-Weighted Governance
May 19, 2026Yuzhe Zhang, Manvir Schneider, Qin Wang +1Majority Voting Governance
Independent researcher, Australia · Cardano Foundation, Switzerland · CSIRO, Australia +1
Diagnosing Korean-Language LLM Political Bias via Census-Grounded Agent Simulation
May 18, 2026Sungwoo KangLarge Language Model Bias Large Language Model Decisions
Department of Electrical and Computer Engineering Korea University, Seoul, Republic of Korea
DuIVRS-2: An LLM-based Interactive Voice Response System for Large-scale POI Attribute Acquisition
May 18, 2026Le Zhang, Shengming Zhang, Rui Zha +3Voice Agents Large Language Model Agents
Baidu Inc., Beijing, China
Edit-level Majority Voting Mitigates Over-Correction in LLM-based Grammatical Error Correction
May 13, 2026Takumi Goto, Yusuke Sakai, Taro WatanabeGrammatical Error Correction Correction
Nara Institute of Science and Technology (NAIST)
The End Justifies the Mean: A Linear Ranking Rule for Proportional Sequential Decisions
May 12, 2026Carmel Baharav, Niclas Boehmer, Bailey Flanigan +1Majority Voting Algorithmic Fairness
MIT, USA · Hasso Plattner Institute, University of Potsdam, Germany
The Price of Proportional Representation in Temporal Voting
May 11, 2026Nicholas TehElections Welfare
University of Oxford
Beyond Majority Voting: Agreement-Based Clustering to Model Annotator Perspectives in Subjective NLP Tasks
May 11, 2026Tadesse Destaw Belay, Ibrahim Said Ahmad, Idris Abdulmumin +6Human Annotators Disagreement
1Instituto Politécnico Nacional · University of Wisconsin–Stevens Point · University of Pretoria +3
The Silent Vote: Improving Zero-Shot LLM Reliability by Aggregating Semantic Neighborhoods
May 10, 2026Sanket Badhe, Priyanka Tiwari, Deep ShahGumbel-Softmax Relaxation Overconfidence
Google
Efficient Ensemble Selection from Binary and Pairwise Feedback
May 10, 2026Tzeh Yuan Neoh, Nicholas Teh, Je Qin Chooi +2Ensemble Submodular Functions
Harvard University, USA · University of Oxford, UK
Semantic Voting: Execution-Grounded Consensus for LLM Code Generation
May 9, 2026Shan Jiang, Zijian Yi, Chenguang ZhuCode Generation Large Language Model Generation
The University of Texas at Austin
VecCISC: Improving Confidence-Informed Self-Consistency with Reasoning Trace Clustering and Candidate Answer Selection
May 8, 2026James Petullo, Sonny George, Dylan Cashman +1Reasoning Traces Majority Voting
Computer Science Department, Brandeis University
Nürnberg NLP at PsyDefDetect: Multi-Axis Voter Ensembles for Psychological Defence Mechanism Classification
May 8, 2026Philipp Steigerwald, Eric Rudolph, Jens AlbrechtPsychology Natural Language Processing
Technische Hochschule Nürnberg Georg Simon Ohm
When Can Voting Help, Hurt, or Change Course? Exact Structure of Binary Test-Time Aggregation
May 7, 2026Yi LiuMajority Voting Exact
York University
Two Calls, Two Moments, and the Vote-Accuracy Curve of Repeated LLM Inference
May 5, 2026Yi LiuMajority Voting LLM Inference Optimization
York University
Instance-Level Costs for Nuanced Classifier Evaluation
May 4, 2026Kabir Kang, Stephen MussmannClassifier Evaluation Metrics
School of Computer Science, Georgia Institute of Technology, Atlanta, GA, USA.
Computing Thiele Rules on Interval Elections and their Generalizations
May 4, 2026Dimitris Avramidis, Alexandra Lassota, Ulrike Schmidt-Kraepelin +1Elections Majority Voting
Department of Mathematics and Computer Science Eindhoven University of Technology Eindhoven, Netherlands · School of Computer Science, and Department of Mathematics and Statistics McGill University Montréal, Canada
SignMuon: Communication-Efficient Distributed Muon Optimization
May 4, 2026Neel Mishra, Kushagara Trivedi, Pawan KumarMuon Decentralized Optimization
1Microsoft · 2IIIT Hyderabad
STABLEVAL: Disagreement-Aware and Stable Evaluation of AI Systems
May 4, 2026Akash Bonagiri, Gerard Janno Anderias, Saee Patil +6Artificial Intelligence Evaluation Disagreement
Department of Computer Science, University of California, Davis, Davis, CA, USA.
The Invisible Coalition Partner: How LLMs Vote When Democracy Gets Concrete
May 3, 2026Joel BarmettlerIdeology Large Language Model Decisions
Independent Researcher Zurich, Switzerland
Fake Plastic Voters: When Political Parties Can Use AI-Simulated Focus Groups
Apr 28, 2026Claudio Novelli, Javier Argota Sanchez-Vaquerizo, Jennifer Cyr +4Political Discourse Scaling Improve Social Simulation
Digital Ethics Center, Yale University, New Haven, CT, USA · Department of Humanities, Social and Political Sciences (GESS), ETH Zürich, Switzerland · Department of Political Science and International Studies, Universidad Torcuato Di Tella, Buenos Aires, Argentina +3
JURY-RL: Votes Propose, Proofs Dispose for Label-Free RLVR
Apr 28, 2026Xinjie Chen, Biao Fu, Jing Wu +4Reinforcement Learning With Verifiable Reward Verifiable Rewards
Tongyi Lab, Alibaba Group, Hangzhou, China
Representational Harms in LLM-Generated Narratives Against Global Majority Nationalities
Apr 24, 2026Ilana Nguyen, Harini Suresh, Thema Monroe-White +1Large Language Model Bias Narratives
Brown University, USA · Brown University, Providence, Rhode Island, USA · George Mason University, Fairfax, Virginia, USA +1
Mapping the Political Discourse in the Brazilian Chamber of Deputies: A Multi-Faceted Computational Approach
Apr 23, 2026Flávio Soriano, Victoria F. Mello, Pedro B. Rigueira +4Parliamentary Debates Political Discourse
Computer Science Department, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil
Is Four Enough? Automated Reasoning Approaches and Dual Bounds for Condorcet Dimensions of Elections
Apr 21, 2026Itai Zilberstein, Ratip Emin Berker, George Li +1Elections Majority Voting
Carnegie Mellon University Pittsburgh, USA
Perturbation Sensitivity of Maximum-Likelihood Pairwise Ranking in Computational Decision Systems
Apr 20, 2026Junyi Yao, Zihao Zheng, Jiayu LongRanking Bounded Adversary
Washington University in St. Louis USA
Learning Unanimously Acceptable Lotteries via Queries
Apr 19, 2026Davin Choo, Paul W. Goldberg, Nicholas TehLotteries Majority Voting
Harvard University, USA · University of Oxford, UK
MCPO: Mastery-Consolidated Policy Optimization for Large Reasoning Models
Apr 18, 2026Zhaokang Liao, Yingguo Gao, Yi Yang +2Reinforcement Learning With Verifiable Reward Frictive Policy Optimization
Ant Group · Zhejiang University
Majority Voting for Code Generation
Apr 17, 2026Tim Launer, Jonas Hübotter, Marco Bagatella +2Code Generation Majority Voting
ETH Zürich, Switzerland
BalLOT: Balanced $k$ -means clustering with optimal transport
Dec 5, 2025Wenyan Luo, Dustin G. MixonK-Means Optimal Transport Approach
Department of Mathematics, The Ohio State University, Columbus, Ohio, USA · Translational Data Analytics Institute, The Ohio State University, Columbus, Ohio, USA
Moloch's Bargain: Emergent Misalignment When LLMs Compete for Audiences
Oct 7, 2025Batu El, James ZouSocial Influence Emergent Misalignment
Stanford University
Algorithmic Shortlisting in Participatory Budgeting
Aug 7, 2025Juan Ignacio Zambrano, Clément Contet, Jairo Gudiño-Rosero +3Majority Voting Democracy
IRIT · Université Toulouse Capitole · Université de Toulouse +3
VOTE: Vision-Language-Action Optimization with Trajectory Ensemble Voting
Jul 7, 2025Juyi Lin, Amir Taherin, Arash Akbari +11Diffusion-Based Vision-Language-Actions Robotic Manipulation
Northeastern University, Boston, USA · EmbodyX,San Mateo,USA
When majority rules, minority loses: bias amplification of gradient descent
May 19, 2025François Bachoc, Jérôme Bolte, Ryan Boustany +1Minority Imbalanced Classification
LPP · University of Lille · Institut Universitaire de France (IUF) +6
Toward Collective-Centric Evaluation of Preference Inference for Participatory Democracy
Date pendingPierre-Antoine Lequeu, Salim Hafid, Paul Lerner +6Democracy Preference Alignment Learning
Sorbonne Université, CNRS, ISIR, Paris, France · Sciences Po, médialab, Paris, France · Sorbonne Université, STIH/CERES, Paris, France
Bias-Corrected Data Synthesis for Imbalanced Learning
Date pendingPengfei Lyu, Zhengchi Ma, Linjun Zhang +1Oversampling Class Imbalance
Department of Biostatistics & Bioinformatics, Duke University · Department of Electrical & Computer Engineering, Duke University · Department of Statistics, Rutgers University +1

May 29, 2026, cs.CLJ/K move · Enter open · S save

Consolidating Rewarded Perturbations for LLM Post-Training

Zheyu Zhang, Shuo Yang, Gjergji Kasneci

Technical University of Munich

Large Language Model Reinforcement Learning Post-Training Majority Voting

Open paper AskPDF arXiv

Post-training of language models is commonly framed as a sample-score-update loop implemented by gradient descent. A recent line of work, exemplified by RandOpt, relocates this loop to weight space, sampling Gaussian perturbations around a pretrained model and ensembling the top-K rewarded specialists at inference. While competitive with PPO and GRPO under matched training compute, this prediction-level ensemble incurs K forward passes per test example and does not extend cleanly to free-form generation. We ask whether the rewarded population can instead be folded into a single deployable model, replacing the inference-time ensemble with one consolidated update. A split-half analysis over 25 model-task pairs reveals reproducible low-rank structure in every case. We turn this geometry into CoRP (Consolidating Rewarded Perturbations), a gradient-free operator that combines reward-weighted aggregation, compatibility-aware reweighting, and a held-out validation gate, with no gradient flowing through the language model. Across five language models from 0.5B to 8B and five tasks covering math, code, and creative writing, CoRP improves the base model by 8.1 points on average. Using one tenth of RandOpt's perturbation budget, CoRP exceeds single-inference RandOpt by 6.5 points and recovers more than half of the gain of the 50-pass majority-vote ensemble, at one forward pass per test example.

Momentum

Related topics

Latest papers 105