On Benchmark Hacking in ML Contests: Modeling, Insights and Design

• URL: http://arxiv.org/abs/2604.22230v1
• Date: Fri, 24 Apr 2026 05:07:28 GMT
• Title: On Benchmark Hacking in ML Contests: Modeling, Insights and Design
• Abstract summary: Benchmark hacking refers to tuning a machine learning model to score highly on certain evaluation criteria without improving true generalization or faithfully solving the intended problem.<n>We study this phenomenon in a generic machine learning contest, where each contestant chooses two types of effort.
• Score: 31.78322449702569
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Benchmark hacking refers to tuning a machine learning model to score highly on certain evaluation criteria without improving true generalization or faithfully solving the intended problem. We study this phenomenon in a generic machine learning contest, where each contestant chooses two types of effort: creative effort that improves model capability as desired by the contest host, and mechanistic effort that only improves the model's fitness to the particular task in contest without contributing to true generalization. We establish the existence of a symmetric monotone pure strategy equilibrium in this competition game. It also provides a natural definition of benchmark hacking in this strategic context by comparing a player's equilibrium effort allocation to that of a single-agent baseline scenario. Under our definition, contestants with types below certain threshold (low types) always engage in benchmark hacking, whereas those above the threshold do not. Furthermore, we show that more skewed reward structures (favoring top-ranked contestants) can elicit more desirable contest outcomes. We also provide empirical evidence to support our theoretical predictions.

Related papers

• Leaderboard Incentives: Model Rankings under Strategic Post-Training [39.30823650986712]
  Influential benchmarks incentivize competing model developers to strategically allocate post-training resources toward improvements on the leaderboard.<n>We show that under mild conditions, a recently proposed evaluation protocol, called tune-before-test, induces a benchmark with a unique Nash equilibrium that ranks models by latent quality.
  arXiv  Detail & Related papers  (2026-03-09T13:33:20Z)
• Pruning as a Game: Equilibrium-Driven Sparsification of Neural Networks [1.1458853556386799]
  We show that sparsity emerges naturally when continued participation becomes a dominated strategy at equilibrium.<n>We derive a simple equilibrium-driven pruning algorithm that jointly updates network parameters and participation variables without relying on explicit importance scores.
  arXiv  Detail & Related papers  (2025-12-26T18:25:38Z)
• LLM Swiss Round: Aggregating Multi-Benchmark Performance via Competitive Swiss-System Dynamics [23.99262273166077]
  Large Language Models (LLMs) and diverse specialized benchmarks require a shift from fragmented, task-specific metrics to a holistic, competitive ranking system.<n>We introduce the novel Competitive Swiss-System Dynamics (CSD) framework, which simulates a sequential contest.<n>CSD provides a more nuanced and context-aware ranking than traditional aggregate scoring and static pairwise models.
  arXiv  Detail & Related papers  (2025-12-24T07:14:31Z)
• Scaling Test-Time Compute to Achieve IOI Gold Medal with Open-Weight Models [72.52332895840279]
  GenCluster is a test-time compute framework that attains IOI gold-level performance using open-weight models.<n>We will show that GenCluster can achieve a gold medal at IOI 2025 for the first time with an open-weight model.
  arXiv  Detail & Related papers  (2025-10-16T02:19:25Z)
• reWordBench: Benchmarking and Improving the Robustness of Reward Models with Transformed Inputs [63.42534498099614]
  We show that state-of-the-art reward models suffer from substantial performance degradation even with minor input transformations.<n>We propose to explicitly train them to assign similar scores to paraphrases, and find that this approach also improves robustness to other distinct kinds of transformations.
  arXiv  Detail & Related papers  (2025-03-14T17:59:41Z)
• A Minimaximalist Approach to Reinforcement Learning from Human Feedback [49.45285664482369]
  We present Self-Play Preference Optimization (SPO), an algorithm for reinforcement learning from human feedback. Our approach is minimalist in that it does not require training a reward model nor unstable adversarial training. We demonstrate that on a suite of continuous control tasks, we are able to learn significantly more efficiently than reward-model based approaches.
  arXiv  Detail & Related papers  (2024-01-08T17:55:02Z)
• Benchmarking Robustness and Generalization in Multi-Agent Systems: A Case Study on Neural MMO [50.58083807719749]
  We present the results of the second Neural MMO challenge, hosted at IJCAI 2022, which received 1600+ submissions. This competition targets robustness and generalization in multi-agent systems. We will open-source our benchmark including the environment wrapper, baselines, a visualization tool, and selected policies for further research.
  arXiv  Detail & Related papers  (2023-08-30T07:16:11Z)
• Improved Bayes Risk Can Yield Reduced Social Welfare Under Competition [99.7047087527422]
  In this work, we demonstrate that competition can fundamentally alter the behavior of machine learning scaling trends. We find many settings where improving data representation quality decreases the overall predictive accuracy across users. At a conceptual level, our work suggests that favorable scaling trends for individual model-providers need not translate to downstream improvements in social welfare.
  arXiv  Detail & Related papers  (2023-06-26T13:06:34Z)
• DualFair: Fair Representation Learning at Both Group and Individual Levels via Contrastive Self-supervision [73.80009454050858]
  This work presents a self-supervised model, called DualFair, that can debias sensitive attributes like gender and race from learned representations. Our model jointly optimize for two fairness criteria - group fairness and counterfactual fairness.
  arXiv  Detail & Related papers  (2023-03-15T07:13:54Z)
• Revisiting the Compositional Generalization Abilities of Neural Sequence Models [23.665350744415004]
  We focus on one-shot primitive generalization as introduced by the popular SCAN benchmark. We demonstrate that modifying the training distribution in simple and intuitive ways enables standard seq-to-seq models to achieve near-perfect generalization performance.
  arXiv  Detail & Related papers  (2022-03-14T18:03:21Z)
• Kaggle forecasting competitions: An overlooked learning opportunity [0.0]
  We review the results from six Kaggle competitions featuring real-life business forecasting tasks. We find that most of the Kaggle datasets are characterized by higher intermittence and entropy than the M-competitions. We find the strong performance of gradient boosted decision trees, increasing success of neural networks for forecasting, and a variety of techniques for adapting machine learning models to the forecasting task.
  arXiv  Detail & Related papers  (2020-09-16T14:14:41Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.