OffsetBias: Leveraging Debiased Data for Tuning Evaluators

• URL: http://arxiv.org/abs/2407.06551v2
• Date: Mon, 7 Oct 2024 08:49:49 GMT
• Title: OffsetBias: Leveraging Debiased Data for Tuning Evaluators
• Authors: Junsoo Park, Seungyeon Jwa, Meiying Ren, Daeyoung Kim, Sanghyuk Choi,
• Abstract summary: We qualitatively identify six types of biases inherent in various judge models. Fine-tuning on our dataset significantly enhances the robustness of judge models against biases.
• Score: 1.5790747258969664
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Employing Large Language Models (LLMs) to assess the quality of generated responses, such as prompting instruct-tuned models or fine-tuning judge models, has become a widely adopted evaluation method. It is also known that such evaluators are vulnerable to biases, such as favoring longer responses. While it is important to overcome this problem, the specifics of these biases remain under-explored. In this work, we qualitatively identify six types of biases inherent in various judge models. We propose EvalBiasBench as a meta-evaluation collection of hand-crafted test cases for each bias type. Additionally, we present de-biasing dataset construction methods and the associated preference dataset OffsetBias. Experimental results demonstrate that fine-tuning on our dataset significantly enhances the robustness of judge models against biases and improves performance across most evaluation scenarios. We release our datasets and the fine-tuned judge model to public.

Related papers

• 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)
• Improving Bias Mitigation through Bias Experts in Natural Language Understanding [10.363406065066538]
  We propose a new debiasing framework that introduces binary classifiers between the auxiliary model and the main model. Our proposed strategy improves the bias identification ability of the auxiliary model.
  arXiv  Detail & Related papers  (2023-12-06T16:15:00Z)
• Fast Model Debias with Machine Unlearning [54.32026474971696]
  Deep neural networks might behave in a biased manner in many real-world scenarios. Existing debiasing methods suffer from high costs in bias labeling or model re-training. We propose a fast model debiasing framework (FMD) which offers an efficient approach to identify, evaluate and remove biases.
  arXiv  Detail & Related papers  (2023-10-19T08:10:57Z)
• Fighting Bias with Bias: Promoting Model Robustness by Amplifying Dataset Biases [5.997909991352044]
  Recent work sought to develop robust, unbiased models by filtering biased examples from training sets. We argue that such filtering can obscure the true capabilities of models to overcome biases. We introduce an evaluation framework defined by a bias-amplified training set and an anti-biased test set.
  arXiv  Detail & Related papers  (2023-05-30T10:10:42Z)
• Think Twice: Measuring the Efficiency of Eliminating Prediction Shortcuts of Question Answering Models [3.9052860539161918]
  We propose a simple method for measuring a scale of models' reliance on any identified spurious feature. We assess the robustness towards a large set of known and newly found prediction biases for various pre-trained models and debiasing methods in Question Answering (QA) We find that while existing debiasing methods can mitigate reliance on a chosen spurious feature, the OOD performance gains of these methods can not be explained by mitigated reliance on biased features.
  arXiv  Detail & Related papers  (2023-05-11T14:35:00Z)
• Balancing Unobserved Confounding with a Few Unbiased Ratings in Debiased Recommendations [4.960902915238239]
  We propose a theoretically guaranteed model-agnostic balancing approach that can be applied to any existing debiasing method. The proposed approach makes full use of unbiased data by alternatively correcting model parameters learned with biased data, and adaptively learning balance coefficients of biased samples for further debiasing.
  arXiv  Detail & Related papers  (2023-04-17T08:56:55Z)
• Systematic Evaluation of Predictive Fairness [60.0947291284978]
  Mitigating bias in training on biased datasets is an important open problem. We examine the performance of various debiasing methods across multiple tasks. We find that data conditions have a strong influence on relative model performance.
  arXiv  Detail & Related papers  (2022-10-17T05:40:13Z)
• General Greedy De-bias Learning [163.65789778416172]
  We propose a General Greedy De-bias learning framework (GGD), which greedily trains the biased models and the base model like gradient descent in functional space. GGD can learn a more robust base model under the settings of both task-specific biased models with prior knowledge and self-ensemble biased model without prior knowledge.
  arXiv  Detail & Related papers  (2021-12-20T14:47:32Z)
• Don't Discard All the Biased Instances: Investigating a Core Assumption in Dataset Bias Mitigation Techniques [19.252319300590656]
  Existing techniques for mitigating dataset bias often leverage a biased model to identify biased instances. The role of these biased instances is then reduced during the training of the main model to enhance its robustness to out-of-distribution data. In this paper, we show that this assumption does not hold in general.
  arXiv  Detail & Related papers  (2021-09-01T10:25:46Z)
• Characterizing Fairness Over the Set of Good Models Under Selective Labels [69.64662540443162]
  We develop a framework for characterizing predictive fairness properties over the set of models that deliver similar overall performance. We provide tractable algorithms to compute the range of attainable group-level predictive disparities. We extend our framework to address the empirically relevant challenge of selectively labelled data.
  arXiv  Detail & Related papers  (2021-01-02T02:11:37Z)
• Improving QA Generalization by Concurrent Modeling of Multiple Biases [61.597362592536896]
  Existing NLP datasets contain various biases that models can easily exploit to achieve high performances on the corresponding evaluation sets. We propose a general framework for improving the performance on both in-domain and out-of-domain datasets by concurrent modeling of multiple biases in the training data. We extensively evaluate our framework on extractive question answering with training data from various domains with multiple biases of different strengths.
  arXiv  Detail & Related papers  (2020-10-07T11:18:49Z)

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.