An Empirical Analysis on Large Language Models in Debate Evaluation
- URL: http://arxiv.org/abs/2406.00050v2
- Date: Tue, 4 Jun 2024 14:51:25 GMT
- Title: An Empirical Analysis on Large Language Models in Debate Evaluation
- Authors: Xinyi Liu, Pinxin Liu, Hangfeng He,
- Abstract summary: We investigate the capabilities and inherent biases of advanced large language models (LLMs) such as GPT-3.5 and GPT-4 in the context of debate evaluation.
We uncover a consistent bias in both GPT-3.5 and GPT-4 towards the second candidate response presented.
We also uncover lexical biases in both GPT-3.5 and GPT-4, especially when label sets carry connotations such as numerical or sequential.
- Score: 10.677407097411768
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: In this study, we investigate the capabilities and inherent biases of advanced large language models (LLMs) such as GPT-3.5 and GPT-4 in the context of debate evaluation. We discover that LLM's performance exceeds humans and surpasses the performance of state-of-the-art methods fine-tuned on extensive datasets in debate evaluation. We additionally explore and analyze biases present in LLMs, including positional bias, lexical bias, order bias, which may affect their evaluative judgments. Our findings reveal a consistent bias in both GPT-3.5 and GPT-4 towards the second candidate response presented, attributed to prompt design. We also uncover lexical biases in both GPT-3.5 and GPT-4, especially when label sets carry connotations such as numerical or sequential, highlighting the critical need for careful label verbalizer selection in prompt design. Additionally, our analysis indicates a tendency of both models to favor the debate's concluding side as the winner, suggesting an end-of-discussion bias.
Related papers
- Evaluating Implicit Bias in Large Language Models by Attacking From a Psychometric Perspective [66.34066553400108]
We conduct a rigorous evaluation of Large Language Models' implicit bias towards certain groups by attacking them with carefully crafted instructions to elicit biased responses.
We propose three attack approaches, i.e., Disguise, Deception, and Teaching, based on which we built evaluation datasets for four common bias types.
arXiv Detail & Related papers (2024-06-20T06:42:08Z) - Unveiling Divergent Inductive Biases of LLMs on Temporal Data [4.561800294155325]
This research focuses on evaluating the performance of GPT-3.5 and GPT-4 models in the analysis of temporal data.
biases toward specific temporal relationships come to light, with GPT-3.5 demonstrating a preference for "AFTER'' in the QA format for both implicit and explicit events, while GPT-4 leans towards "BEFORE''
arXiv Detail & Related papers (2024-04-01T19:56:41Z) - Bias in Language Models: Beyond Trick Tests and Toward RUTEd Evaluation [55.66090768926881]
We study the correspondence between decontextualized "trick tests" and evaluations that are more grounded in Realistic Use and Tangible Effects.
We compare three de-contextualized evaluations adapted from the current literature to three analogous RUTEd evaluations applied to long-form content generation.
We found no correspondence between trick tests and RUTEd evaluations.
arXiv Detail & Related papers (2024-02-20T01:49:15Z) - Revisiting Zero-Shot Abstractive Summarization in the Era of Large Language Models from the Perspective of Position Bias [13.828653029379257]
We characterize and study zero-shot abstractive summarization in Large Language Models (LLMs) by measuring position bias.
Position bias captures the tendency of a model unfairly prioritizing information from certain parts of the input text over others, leading to undesirable behavior.
Our findings lead to novel insights and discussion on performance and position bias of models for zero-shot summarization tasks.
arXiv Detail & Related papers (2024-01-03T21:38:40Z) - GPTBIAS: A Comprehensive Framework for Evaluating Bias in Large Language
Models [83.30078426829627]
Large language models (LLMs) have gained popularity and are being widely adopted by a large user community.
The existing evaluation methods have many constraints, and their results exhibit a limited degree of interpretability.
We propose a bias evaluation framework named GPTBIAS that leverages the high performance of LLMs to assess bias in models.
arXiv Detail & Related papers (2023-12-11T12:02:14Z) - 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) - Large Language Models on Wikipedia-Style Survey Generation: an Evaluation in NLP Concepts [21.150221839202878]
Large Language Models (LLMs) have achieved significant success across various general tasks.
In this work, we examine the proficiency of LLMs in generating succinct survey articles specific to the niche field of NLP in computer science.
We compare both human and GPT-based evaluation scores and provide in-depth analysis.
arXiv Detail & Related papers (2023-08-21T01:32:45Z) - Instructed to Bias: Instruction-Tuned Language Models Exhibit Emergent Cognitive Bias [57.42417061979399]
Recent studies show that instruction tuning (IT) and reinforcement learning from human feedback (RLHF) improve the abilities of large language models (LMs) dramatically.
In this work, we investigate the effect of IT and RLHF on decision making and reasoning in LMs.
Our findings highlight the presence of these biases in various models from the GPT-3, Mistral, and T5 families.
arXiv Detail & Related papers (2023-08-01T01:39:25Z) - A negation detection assessment of GPTs: analysis with the xNot360
dataset [9.165119034384027]
Negation is a fundamental aspect of natural language, playing a critical role in communication and comprehension.
We focus on the identification of negation in natural language using a zero-shot prediction approach applied to our custom xNot360 dataset.
Our findings expose a considerable performance disparity among the GPT models, with GPT-4 surpassing its counterparts and GPT-3.5 displaying a marked performance reduction.
arXiv Detail & Related papers (2023-06-29T02:27:48Z) - News Summarization and Evaluation in the Era of GPT-3 [73.48220043216087]
We study how GPT-3 compares against fine-tuned models trained on large summarization datasets.
We show that not only do humans overwhelmingly prefer GPT-3 summaries, prompted using only a task description, but these also do not suffer from common dataset-specific issues such as poor factuality.
arXiv Detail & Related papers (2022-09-26T01:04:52Z)
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.