CARMO: Dynamic Criteria Generation for Context-Aware Reward Modelling

• URL: http://arxiv.org/abs/2410.21545v2
• Date: Mon, 17 Feb 2025 21:25:09 GMT
• Title: CARMO: Dynamic Criteria Generation for Context-Aware Reward Modelling
• Authors: Taneesh Gupta, Shivam Shandilya, Xuchao Zhang, Rahul Madhavan, Supriyo Ghosh, Chetan Bansal, Huaxiu Yao, Saravan Rajmohan,
• Abstract summary: Reward modeling in large language models is susceptible to reward hacking.<n>We propose Context-Aware Reward Modeling (CARMO) to mitigate this problem.<n>We establish a new state-of-the-art performance in zero-shot settings for generative models, achieving a 2.1% improvement on Reward Bench.
• Score: 27.86204841898399
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Reward modeling in large language models is susceptible to reward hacking, causing models to latch onto superficial features such as the tendency to generate lists or unnecessarily long responses. In reinforcement learning from human feedback (RLHF) and more generally during post-training flawed reward signals often lead to outputs that optimize for these spurious correlates instead of genuine quality or correctness. We propose Context-Aware Reward Modeling (CARMO), a novel approach that first generates dynamic, context-relevant criteria to ground the reward model before producing reward scores. Unlike prior methods that rely on static rubrics, CARMO leverages large language models (LLMs) to adaptively create evaluation criteria such as logical consistency, clarity, and depth tailored to the user query. Our theoretical analysis shows that such criteria generation can mitigate reward hacking. We further demonstrate that CARMO can be distilled into smaller models, reducing the computational cost of alignment. We establish a new state-of-the-art performance in zero-shot settings for generative models, achieving a 2.1\% improvement on Reward Bench. Furthermore, alignment performed on the CARMO-curated preference dataset achieves 22.5\% and 21.1\% LC-WR and WR, respectively, on Mistral-Base (7B).

Related papers

• reWordBench: Benchmarking and Improving the Robustness of Reward Models with Transformed Inputs [64.29893431743608]
  We show that state-of-the-art reward models suffer from substantial performance degradation even with minor input transformations. 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)
• RM-Bench: Benchmarking Reward Models of Language Models with Subtlety and Style [37.97757796124621]
  RM-Bench is a novel benchmark designed to evaluate reward models based on their sensitivity to subtle content differences and resistance to style biases. We evaluate nearly 40 reward models on RM-Bench and find that even state-of-the-art models achieve an average performance of only 46.6%.
  arXiv  Detail & Related papers  (2024-10-21T16:48:26Z)
• Elephant in the Room: Unveiling the Impact of Reward Model Quality in Alignment [50.21842377409232]
  Despite vital role reward models play in alignment, previous works have consistently overlooked their performance. This work first investigates the quality of the widely-used preference dataset, HH-RLHF, and curates a clean version, CHH-RLHF. Based on CHH-RLHF, we benchmark the accuracy of a broad range of reward models used in previous alignment works, unveiling the unreliability of using them both for optimization and evaluation.
  arXiv  Detail & Related papers  (2024-09-26T04:28:35Z)
• Critique-out-Loud Reward Models [20.631830494414096]
  We introduce Critique-out-Loud (CLoud) reward models. CLoud reward models operate by first generating a natural language critique of the assistant's response. We demonstrate the success of CLoud reward models for both Llama-3-8B and 70B base models.
  arXiv  Detail & Related papers  (2024-08-21T17:24:15Z)
• Disce aut Deficere: Evaluating LLMs Proficiency on the INVALSI Italian Benchmark [12.729687989535359]
  evaluating Large Language Models (LLMs) in languages other than English is crucial for ensuring their linguistic versatility, cultural relevance, and applicability in diverse global contexts. We tackle this challenge by introducing a structured benchmark using the INVALSI tests, a set of well-established assessments designed to measure educational competencies across Italy.
  arXiv  Detail & Related papers  (2024-06-25T13:20:08Z)
• RewardBench: Evaluating Reward Models for Language Modeling [100.28366840977966]
  We present RewardBench, a benchmark dataset and code-base for evaluation of reward models. The dataset is a collection of prompt-chosen-rejected trios spanning chat, reasoning, and safety. On the RewardBench leaderboard, we evaluate reward models trained with a variety of methods.
  arXiv  Detail & Related papers  (2024-03-20T17:49:54Z)
• HD-Eval: Aligning Large Language Model Evaluators Through Hierarchical Criteria Decomposition [92.17397504834825]
  HD-Eval is a framework that iteratively aligns large language models evaluators with human preference. HD-Eval inherits the essence from the evaluation mindset of human experts and enhances the alignment of LLM-based evaluators. Extensive experiments on three evaluation domains demonstrate the superiority of HD-Eval in further aligning state-of-the-art evaluators.
  arXiv  Detail & Related papers  (2024-02-24T08:01:32Z)
• Can Large Language Models be Trusted for Evaluation? Scalable Meta-Evaluation of LLMs as Evaluators via Agent Debate [74.06294042304415]
  We propose ScaleEval, an agent-debate-assisted meta-evaluation framework. We release the code for our framework, which is publicly available on GitHub.
  arXiv  Detail & Related papers  (2024-01-30T07:03:32Z)
• F-Eval: Assessing Fundamental Abilities with Refined Evaluation Methods [102.98899881389211]
  We propose F-Eval, a bilingual evaluation benchmark to evaluate the fundamental abilities, including expression, commonsense and logic. For reference-free subjective tasks, we devise new evaluation methods, serving as alternatives to scoring by API models.
  arXiv  Detail & Related papers  (2024-01-26T13:55:32Z)
• Secrets of RLHF in Large Language Models Part II: Reward Modeling [134.97964938009588]
  We introduce a series of novel methods to mitigate the influence of incorrect and ambiguous preferences in the dataset. We also introduce contrastive learning to enhance the ability of reward models to distinguish between chosen and rejected responses.
  arXiv  Detail & Related papers  (2024-01-11T17:56:59Z)
• QualEval: Qualitative Evaluation for Model Improvement [82.73561470966658]
  We propose QualEval, which augments quantitative scalar metrics with automated qualitative evaluation as a vehicle for model improvement. QualEval uses a powerful LLM reasoner and our novel flexible linear programming solver to generate human-readable insights. We demonstrate that leveraging its insights, for example, improves the absolute performance of the Llama 2 model by up to 15% points relative.
  arXiv  Detail & Related papers  (2023-11-06T00:21:44Z)
• 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)
• Calibrating LLM-Based Evaluator [92.17397504834825]
  We propose AutoCalibrate, a multi-stage, gradient-free approach to calibrate and align an LLM-based evaluator toward human preference. Instead of explicitly modeling human preferences, we first implicitly encompass them within a set of human labels. Our experiments on multiple text quality evaluation datasets illustrate a significant improvement in correlation with expert evaluation through calibration.
  arXiv  Detail & Related papers  (2023-09-23T08:46:11Z)
• FLASK: Fine-grained Language Model Evaluation based on Alignment Skill Sets [69.91340332545094]
  We introduce FLASK, a fine-grained evaluation protocol for both human-based and model-based evaluation. We experimentally observe that the fine-graininess of evaluation is crucial for attaining a holistic view of model performance.
  arXiv  Detail & Related papers  (2023-07-20T14:56:35Z)
• Dynamic Model Pruning with Feedback [64.019079257231]
  We propose a novel model compression method that generates a sparse trained model without additional overhead. We evaluate our method on CIFAR-10 and ImageNet, and show that the obtained sparse models can reach the state-of-the-art performance of dense models.
  arXiv  Detail & Related papers  (2020-06-12T15:07:08Z)

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.