Related papers: The Comparative Trap: Pairwise Comparisons Amplifies Biased Preferences of LLM Evaluators

The Comparative Trap: Pairwise Comparisons Amplifies Biased Preferences of LLM Evaluators

URL: http://arxiv.org/abs/2406.12319v3
Date: Wed, 16 Oct 2024 13:39:49 GMT
Title: The Comparative Trap: Pairwise Comparisons Amplifies Biased Preferences of LLM Evaluators
Authors: Hawon Jeong, ChaeHun Park, Jimin Hong, Hojoon Lee, Jaegul Choo,
Abstract summary: Large language models (LLMs) are increasingly used as evaluators for natural language generation tasks. LLMs display biased preferences, such as favoring verbosity and authoritative tones. We introduce PRePair, which integrates pointwise reasoning within a pairwise framework.
Score: 31.520403357740317
License:
Abstract: As large language models (LLMs) are increasingly used as evaluators for natural language generation tasks, ensuring unbiased assessments is essential. However, LLM evaluators often display biased preferences, such as favoring verbosity and authoritative tones. Our empirical analysis reveals that these biases are exacerbated in pairwise evaluation, where LLMs directly compare two outputs and easily prioritize superficial attributes. In contrast, pointwise evaluation, which assesses outputs independently, is less susceptible to such bias because each output is judged in isolation. To address the limitations of the pairwise evaluation, we introduce a novel evaluation method, PRePair, which integrates pointwise reasoning within a pairwise framework. PRePair effectively alleviates biased preference, improving performance on the adversarial benchmark (LLMBar) while outperforming pointwise evaluation on the standard benchmark (MT-Bench).

Related papers

Language Model Preference Evaluation with Multiple Weak Evaluators [78.53743237977677]
GED (Preference Graph Ensemble and Denoise) is a novel approach that leverages multiple model-based evaluators to construct preference graphs. We show that GED outperforms baseline methods in model ranking, response selection, and model alignment tasks.
arXiv Detail & Related papers (2024-10-14T01:57:25Z)
Compare without Despair: Reliable Preference Evaluation with Generation Separability [20.50638483427141]
We introduce a measure, separability, which estimates how suitable a test instance is for pairwise preference evaluation. For a candidate test instance, separability samples multiple generations from a pair of models, and measures how distinguishable the two sets of generations are. Experiments show that instances with high separability values yield more consistent preference ratings from both human- and auto-raters.
arXiv Detail & Related papers (2024-07-02T01:37:56Z)
Fairer Preferences Elicit Improved Human-Aligned Large Language Model Judgments [41.25558612970942]
We show that large language models (LLMs) exhibit preference biases and worrying sensitivity to prompt designs. Motivated by this phenomenon, we propose an automatic Zero-shot Evaluation-oriented Prompt Optimization framework, ZEPO.
arXiv Detail & Related papers (2024-06-17T09:48:53Z)
CASPR: Automated Evaluation Metric for Contrastive Summarization [4.310460539747285]
We propose an automated evaluation metric CASPR to better measure contrast between a pair of summaries. Our results on a prior dataset CoCoTRIP demonstrate that CASPR can more reliably capture the contrastiveness of the summary pairs compared to the baselines.
arXiv Detail & Related papers (2024-04-23T23:27:29Z)
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)
Aligning with Human Judgement: The Role of Pairwise Preference in Large Language Model Evaluators [48.54465599914978]
Large Language Models (LLMs) have demonstrated promising capabilities in assessing the quality of generated natural language. LLMs still exhibit biases in evaluation and often struggle to generate coherent evaluations that align with human assessments. We introduce Pairwise-preference Search (PairS), an uncertainty-guided search method that employs LLMs to conduct pairwise comparisons and efficiently ranks candidate texts.
arXiv Detail & Related papers (2024-03-25T17:11:28Z)
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)
LLM Comparative Assessment: Zero-shot NLG Evaluation through Pairwise Comparisons using Large Language Models [55.60306377044225]
Large language models (LLMs) have enabled impressive zero-shot capabilities across various natural language tasks. This paper explores two options for exploiting the emergent abilities of LLMs for zero-shot NLG assessment. For moderate-sized open-source LLMs, such as FlanT5 and Llama2-chat, comparative assessment is superior to prompt scoring.
arXiv Detail & Related papers (2023-07-15T22:02:12Z)
LLMs as Factual Reasoners: Insights from Existing Benchmarks and Beyond [135.8013388183257]
We propose a new protocol for inconsistency detection benchmark creation and implement it in a 10-domain benchmark called SummEdits. Most LLMs struggle on SummEdits, with performance close to random chance. The best-performing model, GPT-4, is still 8% below estimated human performance.
arXiv Detail & Related papers (2023-05-23T21:50:06Z)

This list is automatically generated from the titles and abstracts of the papers in this site.