Language Model Preference Evaluation with Multiple Weak Evaluators

• URL: http://arxiv.org/abs/2410.12869v4
• Date: Thu, 30 Oct 2025 00:34:12 GMT
• Title: Language Model Preference Evaluation with Multiple Weak Evaluators
• Authors: Zhengyu Hu, Jieyu Zhang, Zhihan Xiong, Alexander Ratner, Kaize Ding, Ranjay Krishna,
• Abstract summary: We introduce PGED, a novel approach that leverages multiple model-based evaluators to construct preference graphs, and then ensembles and denoises these graphs for acyclic, non-contradictory evaluation results.<n>We demonstrate PGED's superiority in three applications: 1) model ranking for evaluation, 2) response selection for test-time scaling, and 3) data selection for model fine-tuning.
• Score: 89.90733463933431
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Despite the remarkable success of Large Language Models (LLMs), evaluating their outputs' quality regarding preference remains a critical challenge. While existing works usually leverage a strong LLM as the judge for comparing LLMs' response pairwisely, such a single-evaluator approach is vulnerable to cyclic preference, i.e., output A is better than B, B than C, but C is better than A, causing contradictory evaluation results. To address this, we introduce PGED (Preference Graph Ensemble and Denoise), a novel approach that leverages multiple model-based evaluators to construct preference graphs, and then ensembles and denoises these graphs for acyclic, non-contradictory evaluation results. We provide theoretical guarantees for our framework, demonstrating its efficacy in recovering the ground truth preference structure. Extensive experiments on ten benchmarks demonstrate PGED 's superiority in three applications: 1) model ranking for evaluation, 2) response selection for test-time scaling, and 3) data selection for model fine-tuning. Notably, PGED combines small LLM evaluators (e.g., Llama3-8B, Mistral-7B, Qwen2-7B) to outperform strong ones (e.g., Qwen2-72B), showcasing its effectiveness in enhancing evaluation reliability and improving model performance.

Related papers

• K-Sort Eval: Efficient Preference Evaluation for Visual Generation via Corrected VLM-as-a-Judge [51.93484138861584]
  The rapid development of visual generative models raises the need for more scalable and human-aligned evaluation methods.<n>We propose K-Sort Eval, a reliable and efficient VLM-based evaluation framework that integrates posterior correction and dynamic matching.<n>Experiments show that K-Sort Eval delivers evaluation results consistent with K-Sort Arena, typically requiring fewer than 90 model runs.
  arXiv  Detail & Related papers  (2026-02-10T05:07:46Z)
• EffiReason-Bench: A Unified Benchmark for Evaluating and Advancing Efficient Reasoning in Large Language Models [32.041688648831794]
  We introduce EffiReason-Bench, a unified benchmark for rigorous cross-paradigm evaluation of efficient reasoning methods.<n>To enable step-by-step evaluation, we construct verified CoT annotations for CommonsenseQA and LogiQA.<n>We propose the E3-Score, a principled metric inspired by economic trade-off modeling that provides smooth, stable evaluation without discontinuities.
  arXiv  Detail & Related papers  (2025-11-13T11:14:46Z)
• Explicit Reasoning Makes Better Judges: A Systematic Study on Accuracy, Efficiency, and Robustness [12.513874407270142]
  We present a systematic comparison of "thinking" and "non-thinking" Large Language Models (LLMs)<n>We evaluate both accuracy and computational efficiency (FLOPs) on RewardBench tasks.<n>Our results show that thinking models achieve approximately 10% points higher accuracy with little overhead.
  arXiv  Detail & Related papers  (2025-09-09T18:36:02Z)
• CPO: Addressing Reward Ambiguity in Role-playing Dialogue via Comparative Policy Optimization [53.79487826635141]
  Reinforcement Learning Fine-Tuning (RLFT) has achieved notable success in tasks with objectively verifiable answers.<n>But it struggles with open-ended subjective tasks like role-playing dialogue.<n>Traditional reward modeling approaches, which rely on independent sample-wise scoring, face dual challenges: subjective evaluation criteria and unstable reward signals.<n>Motivated by the insight that human evaluation inherently combines explicit criteria with implicit comparative judgments, we propose Comparative Policy Optimization.
  arXiv  Detail & Related papers  (2025-08-12T16:49:18Z)
• Rethinking Reward Model Evaluation Through the Lens of Reward Overoptimization [15.729285736811383]
  Reward models play a crucial role in reinforcement learning from human feedback.<n>Existing benchmarks for reward models show a weak correlation with the performance of optimized policies.
  arXiv  Detail & Related papers  (2025-05-19T06:43:08Z)
• From Rankings to Insights: Evaluation Should Shift Focus from Leaderboard to Feedback [36.68929551237421]
  We introduce bftextFeedbacker, an evaluation framework that provides comprehensive and fine-grained results.<n>Our project homepage and dataset are available at https://liudan193.io/Feedbacker.
  arXiv  Detail & Related papers  (2025-05-10T16:52:40Z)
• Pairwise or Pointwise? Evaluating Feedback Protocols for Bias in LLM-Based Evaluation [57.380464382910375]
  We show that the choice of feedback protocol for evaluation can significantly affect evaluation reliability and induce systematic biases.<n>We find that generator models can flip preferences by embedding distractor features.<n>We offer recommendations for choosing feedback protocols based on dataset characteristics and evaluation objectives.
  arXiv  Detail & Related papers  (2025-04-20T19:05:59Z)
• Do LLM Evaluators Prefer Themselves for a Reason? [21.730128682888168]
  Large language models (LLMs) are increasingly used as automatic evaluators in applications such as benchmarking, reward modeling, and self-refinement. Prior work highlights a potential self-preference bias where LLMs favor their own generated responses. This raises a critical question: Is self-preference detrimental, or does it simply reflect objectively superior outputs from more capable models?
  arXiv  Detail & Related papers  (2025-04-04T18:09:23Z)
• Language Models are Few-Shot Graders [0.12289361708127876]
  We present an ASAG pipeline leveraging state-of-the-art LLMs. We compare the grading performance of three OpenAI models: GPT-4, GPT-4o, and o1-preview. Our findings indicate that providing graded examples enhances grading accuracy, with RAG-based selection outperforming random selection.
  arXiv  Detail & Related papers  (2025-02-18T23:38:21Z)
• Offline Model-Based Optimization by Learning to Rank [26.21886715050762]
  We argue that regression models trained with mean squared error (MSE) are not well-aligned with the primary goal of offline model-based optimization. We propose learning a ranking-based model that leverages learning to rank techniques to prioritize promising designs based on their relative scores.
  arXiv  Detail & Related papers  (2024-10-15T11:15:03Z)
• Direct Judgement Preference Optimization [66.83088028268318]
  We train large language models (LLMs) as generative judges to evaluate and critique other models' outputs. We employ three approaches to collect the preference pairs for different use cases, each aimed at improving our generative judge from a different perspective. Our model robustly counters inherent biases such as position and length bias, flexibly adapts to any evaluation protocol specified by practitioners, and provides helpful language feedback for improving downstream generator models.
  arXiv  Detail & Related papers  (2024-09-23T02:08:20Z)
• Self-Taught Evaluators [77.92610887220594]
  We present an approach that aims to im-proves without human annotations, using synthetic training data only. Our Self-Taught Evaluator can improve a strong LLM from 75.4 to 88.3 on RewardBench.
  arXiv  Detail & Related papers  (2024-08-05T17:57:02Z)
• An Optimism-based Approach to Online Evaluation of Generative Models [23.91197677628145]
  We propose an online evaluation framework to find the generative model that maximizes a standard assessment score among a group of available models. Specifically, we study the online assessment of generative models based on the Fr'echet Inception Distance (FID) and Inception Score (IS) metrics.
  arXiv  Detail & Related papers  (2024-06-11T16:57:48Z)
• 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)
• CritiqueLLM: Towards an Informative Critique Generation Model for Evaluation of Large Language Model Generation [87.44350003888646]
  Eval-Instruct can acquire pointwise grading critiques with pseudo references and revise these critiques via multi-path prompting. CritiqueLLM is empirically shown to outperform ChatGPT and all the open-source baselines.
  arXiv  Detail & Related papers  (2023-11-30T16:52:42Z)
• Don't Make Your LLM an Evaluation Benchmark Cheater [142.24553056600627]
  Large language models(LLMs) have greatly advanced the frontiers of artificial intelligence, attaining remarkable improvement in model capacity. To assess the model performance, a typical approach is to construct evaluation benchmarks for measuring the ability level of LLMs. We discuss the potential risk and impact of inappropriately using evaluation benchmarks and misleadingly interpreting the evaluation results.
  arXiv  Detail & Related papers  (2023-11-03T14:59:54Z)
• PRD: Peer Rank and Discussion Improve Large Language Model based Evaluations [10.709365940160685]
  Modern large language models (LLMs) are hard to evaluate and compare automatically. We propose a peer rank (PR) algorithm that takes into account each peer LLM's pairwise preferences of all answer pairs. We find that our approaches achieve higher accuracy and align better with human judgments.
  arXiv  Detail & Related papers  (2023-07-06T04:05:44Z)
• GREAT Score: Global Robustness Evaluation of Adversarial Perturbation using Generative Models [60.48306899271866]
  We present a new framework, called GREAT Score, for global robustness evaluation of adversarial perturbation using generative models. We show high correlation and significantly reduced cost of GREAT Score when compared to the attack-based model ranking on RobustBench. GREAT Score can be used for remote auditing of privacy-sensitive black-box models.
  arXiv  Detail & Related papers  (2023-04-19T14:58:27Z)

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.