Rethinking LLM-based Preference Evaluation
- URL: http://arxiv.org/abs/2407.01085v1
- Date: Mon, 1 Jul 2024 08:37:41 GMT
- Title: Rethinking LLM-based Preference Evaluation
- Authors: Zhengyu Hu, Linxin Song, Jieyu Zhang, Zheyuan Xiao, Jingang Wang, Zhenyu Chen, Jieyu Zhao, Hui Xiong,
- Abstract summary: Large language model (LLM)-based preference evaluation has been widely adopted to compare pairs of model responses.
A severe bias towards lengthy responses has been observed, raising concerns about the reliability of this evaluation method.
- Score: 38.62663118795261
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Recently, large language model (LLM)-based preference evaluation has been widely adopted to compare pairs of model responses. However, a severe bias towards lengthy responses has been observed, raising concerns about the reliability of this evaluation method. In this work, we designed a series of controlled experiments to study the major impacting factors of the metric of LLM-based preference evaluation, i.e., win rate, and conclude that the win rate is affected by two axes of model response: desirability and information mass, where the former is length-independent and related to trustworthiness, and the latter is length-dependent and can be represented by conditional entropy. We find that length impacts the existing evaluations by influencing information mass. However, a reliable evaluation metric should not only assess content quality but also ensure that the assessment is not confounded by extraneous factors such as response length. Therefore, we propose a simple yet effective adjustment, AdapAlpaca, to the existing practice of win rate measurement. Specifically, by adjusting the lengths of reference answers to match the test model's answers within the same interval, we debias information mass relative to length, ensuring a fair model evaluation.
Related papers
- Uncertainty Estimation of Large Language Models in Medical Question Answering [60.72223137560633]
Large Language Models (LLMs) show promise for natural language generation in healthcare, but risk hallucinating factually incorrect information.
We benchmark popular uncertainty estimation (UE) methods with different model sizes on medical question-answering datasets.
Our results show that current approaches generally perform poorly in this domain, highlighting the challenge of UE for medical applications.
arXiv Detail & Related papers (2024-07-11T16:51:33Z) - Challenges and Considerations in the Evaluation of Bayesian Causal Discovery [49.0053848090947]
Representing uncertainty in causal discovery is a crucial component for experimental design, and more broadly, for safe and reliable causal decision making.
Unlike non-Bayesian causal discovery, which relies on a single estimated causal graph and model parameters for assessment, causal discovery presents challenges due to the nature of its quantity.
No consensus on the most suitable metric for evaluation.
arXiv Detail & Related papers (2024-06-05T12:45:23Z) - Mitigating LLM Hallucinations via Conformal Abstention [70.83870602967625]
We develop a principled procedure for determining when a large language model should abstain from responding in a general domain.
We leverage conformal prediction techniques to develop an abstention procedure that benefits from rigorous theoretical guarantees on the hallucination rate (error rate)
Experimentally, our resulting conformal abstention method reliably bounds the hallucination rate on various closed-book, open-domain generative question answering datasets.
arXiv Detail & Related papers (2024-04-04T11:32:03Z) - One Prompt To Rule Them All: LLMs for Opinion Summary Evaluation [30.674896082482476]
We show that Op-I-Prompt emerges as a good alternative for evaluating opinion summaries achieving an average Spearman correlation of 0.70 with humans.
To the best of our knowledge, we are the first to investigate LLMs as evaluators on both closed-source and open-source models in the opinion summarization domain.
arXiv Detail & Related papers (2024-02-18T19:13:52Z) - Linked shrinkage to improve estimation of interaction effects in
regression models [0.0]
We develop an estimator that adapts well to two-way interaction terms in a regression model.
We evaluate the potential of the model for inference, which is notoriously hard for selection strategies.
Our models can be very competitive to a more advanced machine learner, like random forest, even for fairly large sample sizes.
arXiv Detail & Related papers (2023-09-25T10:03:39Z) - REAM$\sharp$: An Enhancement Approach to Reference-based Evaluation
Metrics for Open-domain Dialog Generation [63.46331073232526]
We present an enhancement approach to Reference-based EvAluation Metrics for open-domain dialogue systems.
A prediction model is designed to estimate the reliability of the given reference set.
We show how its predicted results can be helpful to augment the reference set, and thus improve the reliability of the metric.
arXiv Detail & Related papers (2021-05-30T10:04:13Z) - A Statistical Analysis of Summarization Evaluation Metrics using
Resampling Methods [60.04142561088524]
We find that the confidence intervals are rather wide, demonstrating high uncertainty in how reliable automatic metrics truly are.
Although many metrics fail to show statistical improvements over ROUGE, two recent works, QAEval and BERTScore, do in some evaluation settings.
arXiv Detail & Related papers (2021-03-31T18:28:14Z) - SAFEval: Summarization Asks for Fact-based Evaluation [40.02686002117778]
We extend previous approaches and propose a unified framework, named SAFEval.
In contrast to established metrics such as ROUGE or BERTScore, SAFEval does not require any ground-truth reference.
We show that SAFEval substantially improves the correlation with human judgments over four evaluation dimensions.
arXiv Detail & Related papers (2021-03-23T17:16:09Z) - User and Item-aware Estimation of Review Helpfulness [4.640835690336653]
We investigate the role of deviations in the properties of reviews as helpfulness determinants.
We propose a novel helpfulness estimation model that extends previous ones.
Our model is thus an effective tool to select relevant user feedback for decision-making.
arXiv Detail & Related papers (2020-11-20T15:35:56Z)
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.