Pretraining on the Test Set Is No Longer All You Need: A Debate-Driven Approach to QA Benchmarks

• URL: http://arxiv.org/abs/2507.17747v1
• Date: Wed, 23 Jul 2025 17:58:14 GMT
• Title: Pretraining on the Test Set Is No Longer All You Need: A Debate-Driven Approach to QA Benchmarks
• Authors: Linbo Cao, Jinman Zhao,
• Abstract summary: We propose a debate-driven evaluation paradigm that transforms any existing QA dataset into structured adversarial debates.<n>We make two main contributions: (1) an evaluation pipeline to systematically convert QA tasks into debate-based assessments, and (2) a public benchmark that demonstrates our paradigm's effectiveness on a subset of MMLU-Pro questions.
• Score: 2.3188831772813105
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: As frontier language models increasingly saturate standard QA benchmarks, concerns about data contamination, memorization, and escalating dataset creation costs persist. We propose a debate-driven evaluation paradigm that transforms any existing QA dataset into structured adversarial debates--where one model is given the official answer to defend, and another constructs and defends an alternative answer--adjudicated by a judge model blind to the correct solution. By forcing multi-round argumentation, this approach substantially increases difficulty while penalizing shallow memorization, yet reuses QA items to reduce curation overhead. We make two main contributions: (1) an evaluation pipeline to systematically convert QA tasks into debate-based assessments, and (2) a public benchmark that demonstrates our paradigm's effectiveness on a subset of MMLU-Pro questions, complete with standardized protocols and reference models. Empirical results validate the robustness of the method and its effectiveness against data contamination--a Llama 3.1 model fine-tuned on test questions showed dramatic accuracy improvements (50% -> 82%) but performed worse in debates. Results also show that even weaker judges can reliably differentiate stronger debaters, highlighting how debate-based evaluation can scale to future, more capable systems while maintaining a fraction of the cost of creating new benchmarks. Overall, our framework underscores that "pretraining on the test set is no longer all you need," offering a sustainable path for measuring the genuine reasoning ability of advanced language models.

Related papers

• InspireDebate: Multi-Dimensional Subjective-Objective Evaluation-Guided Reasoning and Optimization for Debating [15.096294311783836]
  Existing large language models (LLMs) focus on responding to specific arguments while neglecting objective assessments such as authenticity and logical validity.<n>We propose a dual-component framework: $textbfInspireScore$, a novel evaluation system, and $textbfInspireDebate$, an optimized debating framework.<n>$textbfInspireScore$ achieves 44$%$ higher correlation with expert judgments compared to existing methods, while $textbfInspireDebate$ shows significant improvements.
  arXiv  Detail & Related papers  (2025-06-22T17:14:29Z)
• ICPC-Eval: Probing the Frontiers of LLM Reasoning with Competitive Programming Contests [85.72404266850982]
  We propose textbfICPC-Eval, a top-level competitive coding benchmark designed to probing the frontiers of reasoning.<n>ICPC-Eval includes 118 carefully curated problems from 11 recent ICPC contests held in various regions of the world.<n>Results underscore the significant challenge in evaluating complex reasoning abilities.
  arXiv  Detail & Related papers  (2025-06-05T11:20:37Z)
• Q-Ponder: A Unified Training Pipeline for Reasoning-based Visual Quality Assessment [10.701522670464463]
  multimodal large language models (MLLMs) can proficiently evaluate visual quality through interpretable assessments.<n>We propose a unified two-stage training framework comprising a cold-start stage and a reinforcement learning-based fine-tuning stage.<n>We designate the models derived from these two stages as Q-Ponder-CI and Q-Ponder.
  arXiv  Detail & Related papers  (2025-06-03T10:11:51Z)
• FINEREASON: Evaluating and Improving LLMs' Deliberate Reasoning through Reflective Puzzle Solving [90.88021670297664]
  FINEREASON is a logic-puzzle benchmark for evaluation of large language models' reasoning capabilities.<n>We introduce two tasks: state checking, and state transition, for a comprehensive evaluation of how models assess the current situation and plan the next move.<n>We show that models trained on our state checking and transition data demonstrate gains in math reasoning by up to 5.1% on GSM8K.
  arXiv  Detail & Related papers  (2025-02-27T16:23:25Z)
• ConQRet: Benchmarking Fine-Grained Evaluation of Retrieval Augmented Argumentation with LLM Judges [23.179246872272362]
  Computational argumentation has become increasingly important in today's polarized environment.<n>We propose a new benchmark featuring long and complex human-authored arguments on debated topics, grounded in real-world websites.<n>Our proposed LLM Judges and the ConQRet benchmark can enable rapid progress in computational argumentation.
  arXiv  Detail & Related papers  (2024-12-06T17:35:52Z)
• Towards Robust Extractive Question Answering Models: Rethinking the Training Methodology [0.34530027457862006]
  Previous research has shown that existing models, when trained on EQA datasets that include unanswerable questions, demonstrate a significant lack of robustness. Our proposed training method includes a novel loss function for the EQA problem and challenges an implicit assumption present in numerous EQA datasets. Our models exhibit significantly enhanced robustness against two types of adversarial attacks, with a performance decrease of only about a third compared to the default models.
  arXiv  Detail & Related papers  (2024-09-29T20:35:57Z)
• Training Language Models to Win Debates with Self-Play Improves Judge Accuracy [8.13173791334223]
  We test the robustness of debate as a method of scalable oversight by training models to debate with data generated via self-play. We find that language model based evaluators answer questions more accurately when judging models optimized to win debates.
  arXiv  Detail & Related papers  (2024-09-25T05:28:33Z)
• MR-Ben: A Meta-Reasoning Benchmark for Evaluating System-2 Thinking in LLMs [55.20845457594977]
  Large language models (LLMs) have shown increasing capability in problem-solving and decision-making.<n>We present a process-based benchmark MR-Ben that demands a meta-reasoning skill.<n>Our meta-reasoning paradigm is especially suited for system-2 slow thinking.
  arXiv  Detail & Related papers  (2024-06-20T03:50:23Z)
• Detecting Multimodal Situations with Insufficient Context and Abstaining from Baseless Predictions [75.45274978665684]
  Vision-Language Understanding (VLU) benchmarks contain samples where answers rely on assumptions unsupported by the provided context.<n>We collect contextual data for each sample whenever available and train a context selection module to facilitate evidence-based model predictions.<n>We develop a general-purpose Context-AwaRe Abstention detector to identify samples lacking sufficient context and enhance model accuracy.
  arXiv  Detail & Related papers  (2024-05-18T02:21:32Z)
• Evaluating Generative Language Models in Information Extraction as Subjective Question Correction [49.729908337372436]
  We propose a new evaluation method, SQC-Score. Inspired by the principles in subjective question correction, we propose a new evaluation method, SQC-Score. Results on three information extraction tasks show that SQC-Score is more preferred by human annotators than the baseline metrics.
  arXiv  Detail & Related papers  (2024-04-04T15:36:53Z)
• Improving Machine Translation with Human Feedback: An Exploration of Quality Estimation as a Reward Model [75.66013048128302]
  In this work, we investigate the potential of employing the QE model as the reward model to predict human preferences for feedback training. We first identify the overoptimization problem during QE-based feedback training, manifested as an increase in reward while translation quality declines. To address the problem, we adopt a simple yet effective method that uses rules to detect the incorrect translations and assigns a penalty term to the reward scores of them.
  arXiv  Detail & Related papers  (2024-01-23T16:07:43Z)
• Position: AI Evaluation Should Learn from How We Test Humans [65.36614996495983]
  We argue that psychometrics, a theory originating in the 20th century for human assessment, could be a powerful solution to the challenges in today's AI evaluations.
  arXiv  Detail & Related papers  (2023-06-18T09:54:33Z)

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.