Correct

Follow this topic Ask about this topic

Momentum

10 papers in the last four weeks, up 150% on the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 32

All topics

From Discovery to Decision: Finite-Budget Recoverability in LLM Voting
Oct 1, 2026Shaoang Li, Jian LiMajority Voting Large Language Model Responses
Stony Brook University
Right Answer, Wrong Mechanism: Detecting Pernicious Divergence in Causal Interventions
Sep 30, 2026Beiming Liu, Minjie ChenCausal Intervention Divergence
PetroChina Southwest Oil and Gasfield Company, China
HARISSA: Inference-Time Self-Checks for Efficient and Safe Local Language Model Deployment
Sep 29, 2026Kenan Alkiek, Moontae Lee, David Jurgens +1Inference-Time Prefill
School of Information University of Michigan Ann Arbor, MI 48109, USA · SNU-LG AI Research Center Seoul, South Korea
Verifier Errors in RLVR: Reward Hacking, Limits of Feedback, and Selective Control
Sep 28, 2026Christian Moya, Elliott Thornley, Guang LinReinforcement Learning With Verifiable Reward Verifiable Rewards
Purdue University · National University of Singapore
Correct then Forecast: Observer State-Space Models for Time Series Forecasting
Sep 27, 2026Alexis-Raja Brachet, Guillaume Clavier--Frémond, Abdelhakim Ziani +2State Space Models Time Series
MICS, CentraleSup´elec, Universit´e Paris-Saclay, France · CentraleSup´elec, IETR UMR CNRS 6164, France · Universit`a di Torino, Torino, Italy.
A Wrong Turn Does Not Ruin the Journey: Deviation-Guided Skill Self-Evolution for LLM Agents
Sep 24, 2026Yichun Feng, Jiawei Wang, Haozhe SunSkill Evolution Large Language Model Agents
University of Chinese Academy of Sciences · University of Science and Technology of China · Meituan
Silent Failures Beyond the 32-Bit Index Range: A Differential Characterization of Large-Tensor Matrix Multiplication in PyTorch's MPS Backend
Sep 19, 2026Junichiro NiimiMatrix Multiplication Pytorch
Meijo University
Correct Now, Insufficient Later: Auditing Update Sufficiency in Context Compression
Sep 17, 2026Guangzhe ZhangModel Auditing Correct
Independent AI Researcher
A Chosen Future Can Still Be Rewritten: Causal Writability in Video Models
Sep 14, 2026Xingyun Wang, Haomin Zheng, Man Yuan +2Correct
Tsinghua University · MetaCircle · Peking University +2
When Models Defer to Wrong Answers: A Robustness Audit of Source-Attributed Cues in Multiple-Choice QA
Sep 8, 2026Manikandan Ravikiran, Siddharth VohraMultiple-Choice Questions Correct
Indian Institute of Technology Mandi, India · Carnegie Mellon University Amazon Web Services AI Native Pittsburgh, PA, USA
Right Frame, Wrong Rule: Cultural Cues Expose the Financial Knowledge Gap They Were Meant to Close
Sep 1, 2026Rania Elbadry, Ahmed Heakl, Saeed Almheiri +12Social Norms Interpretive Layer
♠MBZUAI · ♢The University of Tokyo · §The Fin AI +4
The Answer Is Not the Argument
Aug 31, 2026Will Yeadon, Sergio Juárez, Paul Mackay +5Answer Reasoning Traces
Department of Physics, Durham University, Durham DH1 3LE, UK · Escuela de Ingenier´ıa de Telecomunicaci´on, Department of Signal Theory and Communications, University of Vigo, Vigo E-36310, Spain · Independent Researcher
Correct Is Not Governed: Provenance Integrity in Agentic Workflows
Aug 13, 2026Jesus SalasAgentic Workflows Accountability
Independent Researcher · The author is employed by Microsoft.
Measuring the Wrong Thing: Internal Harmfulness Scores Anti-Rank Successful Jailbreaks
Aug 10, 2026Mingyu Luo, Ming Deng, Zilang Qiu +8Jailbreak Attacks Attack-Success Rate
College of Computer Science and Artificial Intelligence, Fudan University · School of Computer Engineering and Science, Shanghai University · Beijing Normal University +5
Wrong Operator or Blind Design? A Reference-Free Diagnostic for Physics-Informed Coefficient Learning
Aug 4, 2026Eric FockParametric Physics-Informed Neural Network Physics-Informed Learning
PIMENT Laboratory, Universit´e de La R´eunion, Le Tampon 97430, La R´eunion, France
Every Wrong Answer Counts: Option-Level Psychometrics for LLM Multiple-Choice Benchmarks
Aug 3, 2026Xiao Fei, Yang Zhang, Sarah Almeida Carneiro +1Psychometric Properties Multiple-Choice Questions
École Polytechnique, Institut Polytechnique de Paris, France · Mohamed bin Zayed University of Artificial Intelligence, United Arab Emirates
Wrong Design Intent Is Worse Than Never Conditioning: A Derangement-Control Diagnosis of Header Conditioning in CAD Program Completion
Jul 25, 2026Yang XiaoCorrect Instruction-Tuned Models
Sichuan University
Detectors Learn the Wrong Thing: Shortcut-Resistant Adversarial Training Against Physically Realizable Attacks
Jul 23, 2026Yuanhao Huang, Yilong Ren, Jinlei Wang +3Unsupervised Detection Adversarial Training
School of Transportation Science and Engineering, Beihang University, 100191, Beijing, P.R China · Zhongguancun Laboratory, 100191, Beijing, P.R China · School of Systems Science and Engineering, Sun Yat-Sen University, 510006, Guangzhou, P.R China +2
Silent Failures in Quantized LLM Reasoning: A Taxonomy-Based Analysis of Hollow Convergence and Failure Mode Shifts
Jul 10, 2026Renuka Oladri, Mohan Vamsi Varadaraju Priya, Jerry WuLarge Language Model Quantization Large Language Models Fail
Science Academy University of Maryland College Park, MD, USA · Electrical and Computer Engineering University of Maryland College Park, MD, USA
Driving the Wrong Way: Leveraging Interpretability in End2End Autonomous Driving Models
Jul 7, 2026Franz Motzkus, Sebastian BernhardCausality-Aware End-To-End Autonomous Driving Interpretability
AUMOVIO, Germany · University of Bamberg, Germany
Position: Correct Answer, Wrong Mechanism -- When AI Scientists Defend General Claims Their Own Data Contradicts
Jun 22, 2026Steven Young EuligArtificial Intelligence Scientists Correct
Department of Physics and Laboratory for Particle Physics and Cosmology (LPPC), Harvard University, Cambridge, MA 02138, USA.
When Confidence Takes the Wrong Path: Diagnosing Retrieval-State Lock-In in RAG
Jun 22, 2026Sahib JulkaLock-In Correct
LMU University Hospital, LMU Munich · Marchioninistr. 15, 81377 Munich
The Wrong Kind of Right: Quantifying and Localizing Misfired Alignment in LLMs
Jun 17, 2026Naihao Deng, Yiming Feng, Chimaobi Okite +4Large Language Model Alignment Misalignment Persona
University of Michigan · University of Cambridge · University of Aberdeen
Zone of Proximal Policy Optimization: Teacher in Prompts, Not Gradients
Jun 16, 2026Byung-Kwan Lee, Ximing Lu, Shizhe Diao +6Proximal Policy Optimization Teacher
NVIDIA
Correct Looks Better: Pairwise Comparisons Reveal Accuracy Rankings
Jun 8, 2026Mina Remeli, Moritz HardtSpearman Correlation Correct
Max Planck Institute for Intelligent Systems, Tübingen, Germany, Tübingen AI Center
CEPO: RLVR Self-Distillation using Contrastive Evidence Policy Optimization
May 19, 2026Ahmed Heakl, Abdelrahman M. Shaker, Youssef Mohamed +4Reinforcement Learning With Verifiable Reward Mathematical Reasoning Benchmarks
MBZUAI · Linköping University · Australian National University
Your Simulation Runs but Solves the Wrong Physics: PDE-Grounded Intent Verification for LLM-Generated Multiphysics Simulation Code
May 10, 2026Zhenghan Song, Yulong Liu, Cheng Wan +4Code Generation Physics Simulation
Cornell University · Columbia University · Harvard University & Nanyang Technological University
Correct Is Not Enough: Training Reasoning Planners with Executor-Grounded Rewards
May 5, 2026Tianyang Han, Hengyu Shi, Junjie Hu +3Reasoning Traces Reinforcement Learning With Verifiable Reward
1D4 Lab · 2Independent Researcher
The Right Answer, the Wrong Direction: Why Transformers Fail at Counting and How to Fix It
May 5, 2026Gabriel GarciaTransformer Architectures Counting
Independent Researcher
Perturb and Correct: Post-Hoc Ensembles using Affine Redundancy
May 2, 2026Eleanor QuintPost-Hoc Per-Node Correction
Department of Computing University of Nebraska–Lincoln Lincoln, NE 68508, USA
DART: Mitigating Harm Drift in Difference-Aware LLMs via Distill-Audit-Repair Training
Apr 18, 2026Ziwen Pan, Zihan Liang, Jad Kabbara +1Large Language Model Safety Harms
Emory University · MIT
How Transformers Reject Wrong Answers: Rotational Dynamics of Factual Constraint Processing
Feb 25, 2026Javier MarínTransformer Encoder Transformer Architectures
Groundlens-dev

Oct 1, 2026, cs.AIJ/K move · Enter open · S save

From Discovery to Decision: Finite-Budget Recoverability in LLM Voting

Shaoang Li, Jian Li

Stony Brook University

Majority Voting Large Language Model Responses Correct Decisions Reasoning Gaps Discovery

Open paper AskPDF arXiv

Voting over multiple LLM responses is a common primitive in test-time scaling and ensemble inference. Collecting more responses can expand the candidate pool and increase the chance that a correct answer is discovered. Under a fixed call budget, a discovered answer still needs to accumulate enough support within the remaining calls to become the final plurality winner, creating a discovery-to-decision gap. In this work, we characterize this gap through the realized vote state and remaining call budget. We derive a sharp recoverability threshold and show that, as sampling proceeds, the observed candidate set can only expand while the set of reachable endpoint winners can only contract, inducing a candidate-level conversion window. Under a specified iid response law, the same state yields exact finite-horizon endpoint probabilities. We further show that merging wrong-answer identities preserves single-call correctness and cannot improve plurality accuracy, and that the effect of redistributing wrong-answer probability depends on the realized vote state. Singleton reachability yields a gold-free exact locking certificate. For a known answer universe, its first trigger is the earliest prefix at which all admissible continuations yield the same fixed-budget output. Empirically, most discovered-but-unselected correct answers lose reachability only after discovery. In a controlled Word16 study, input permutation improves raw-plurality accuracy by 21.1 points with essentially unchanged single-call correctness. Exact locking saves 28-30% of calls at a 16-call budget while preserving every fixed-budget output.

Momentum

Related topics

Latest papers 32