An official website of the United States government
This paper investigates the performance of a diverse set of large language models (LLMs) including leading closed-source (GPT-4, GPT-4o mini, Claude 3.5 Haiku) and open-source (Llama 3.1 70B, Llama 3.1 8B) models, alongside the earlier GPT-3.5 within the context of U.S. tax resolutions. AI-driven solutions like these have made substantial inroads into legal-critical systems with significant socio-economic implications. However, their accuracy and reliability have not been assessed in some legal domains, such as tax. Using the Volunteer Income Tax Assistance (VITA) certification tests—endorsed by the US Internal Revenue Service (IRS) for tax volunteering—this study compares these LLMs to evaluate their potential utility in assisting both tax volunteers as well as taxpayers, particularly those with low and moderate income. Since the answers to these questions are not publicly available, we first analyze 130 questions with the tax domain experts and develop the ground truths for each question. We then benchmarked these diverse LLMs against the ground truths using both the original VITA questions and syntactically perturbed versions (a total of 390 questions) to assess genuine understanding versus memorization/hallucinations. Our comparative analysis reveals distinct performance differences: closed-source models (GPT-4, Claude 3.5 Haiku, GPT-4o mini) generally demonstrated higher accuracy and robustness compared to GPT-3.5 and the open-source Llama models. For instance, on basic multiple-choice questions, top models like GPT-4 and Claude 3.5 Haiku achieved 83.33% accuracy, surpassing GPT-3.5 (54.17%) and the open-source Llama 3.1 8B (50.00%). These findings generally hold across both original and perturbed questions. However, the paper acknowledges that these developments are initial indicators, and further research is necessary to fully understand the implications of deploying LLMs in this domain. A critical limitation observed across all evaluated models was significant difficulty with open-ended questions, which require accurate numerical calculation and application of tax rules. We hope that this paper provides a means and a standard to evaluate the efficacy of current and future LLMs in the tax domain.
more »
« less
Characterizing Multimodal Long-form Summarization: A Case Study on Financial Reports
As large language models (LLMs) expand the power of natural language processing to handle long inputs, rigorous and systematic analyses are necessary to understand their abilities and behavior. A salient application is summarization, due to its ubiquity and controversy (e.g., researchers have declared the death of summarization). In this paper, we use financial report summarization as a case study because financial reports are not only long but also use numbers and tables extensively. We propose a computational framework for characterizing multimodal long-form summarization and investigate the behavior of Claude 2.0/2.1, GPT-4/3.5, and Cohere. We find that GPT-3.5 and Cohere fail to perform this summarization task meaningfully. For Claude 2 and GPT-4, we analyze the extractiveness of the summary and identify a position bias in LLMs. This position bias disappears after shuffling the input for Claude, which suggests that Claude seems to recognize important information. We also conduct a comprehensive investigation on the use of numeric data in LLM-generated summaries and offer a taxonomy of numeric hallucination. We employ prompt engineering to improve GPT-4's use of numbers with limited success. Overall, our analyses highlight the strong capability of Claude 2 in handling long multimodal inputs compared to GPT-4. The generated summaries and evaluation code are available at https://github.com/ChicagoHAI/characterizing-multimodal-long-form-summarization.
more »
« less
- PAR ID:
- 10574853
- Publisher / Repository:
- COLM
- Date Published:
- Format(s):
- Medium: X
- Location:
- Philadelphia, PA, USA
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Recent advances in large language models (LLMs) have demonstrated remarkable successes in zero- and few-shot performance on various downstream tasks, paving the way for applications in high-stakes domains. In this study, we systematically examine the capabilities and limitations of LLMs, specifically GPT-3.5 and ChatGPT, in performing zero-shot medical evidence summarization across six clinical domains. We conduct both automatic and human evaluations, covering several dimensions of summary quality. Our study demonstrates that automatic metrics often do not strongly correlate with the quality of summaries. Furthermore, informed by our human evaluations, we define a terminology of error types for medical evidence summarization. Our findings reveal that LLMs could be susceptible to generating factually inconsistent summaries and making overly convincing or uncertain statements, leading to potential harm due to misinformation. Moreover, we find that models struggle to identify the salient information and are more error-prone when summarizing over longer textual contexts.more » « less
-
The evolution of multimodal large language models (LLMs) capable of processing diverse input modalities (e.g., text and images) holds new prospects for their application in engineering design, such as the generation of 3D computer-aided design (CAD) models. However, little is known about the ability of multimodal LLMs to generate 3D design objects, and there is a lack of quantitative assessment. In this study, we develop an approach to enable LLMs to generate 3D CAD models (i.e., LLM4CAD) and perform experiments to evaluate their efficacy where GPT-4 and GPT-4V were employed as examples. To address the challenge of data scarcity for multimodal LLM studies, we created a data synthesis pipeline to generate CAD models, sketches, and image data of typical mechanical components (e.g., gears and springs) and collect their natural language descriptions with dimensional information using Amazon Mechanical Turk. We positioned the CAD program (programming script for CAD design) as a bridge, facilitating the conversion of LLMs’ textual output into tangible CAD design objects. We focus on two critical capabilities: the generation of syntactically correct CAD programs (Cap1) and the accuracy of the parsed 3D shapes (Cap2) quantified by intersection over union. The results show that both GPT-4 and GPT-4V demonstrate great potential in 3D CAD generation by just leveraging their zero-shot learning ability. Specifically, on average, GPT-4V outperforms when processing only text-based input, exceeding the results obtained using multimodal inputs, such as text with image, for Cap 1 and Cap 2. However, when examining category-specific results of mechanical components, the prominence of multimodal inputs is increasingly evident for more complex geometries (e.g., springs and gears) in both Cap 1 and Cap 2. The potential of multimodal LLMs to improve 3D CAD generation is clear, but their application must be carefully calibrated to the complexity of the target CAD models to be generated.more » « less
-
The evolution of multimodal large language models (LLMs) capable of processing diverse input modalities (e.g., text and images) holds new prospects for their application in engineering design, such as the generation of 3D computer-aided design (CAD) models. However, little is known about the ability of multimodal LLMs to generate 3D design objects, and there is a lack of quantitative assessment. In this study, we develop an approach to enable two LLMs, GPT-4 and GPT-4V, to generate 3D CAD models (i.e., LLM4CAD) and perform experiments to evaluate their efficacy. To address the challenge of data scarcity for multimodal LLM studies, we created a data synthesis pipeline to generate CAD models, sketches, and image data of typical mechanical components (e.g., gears and springs) and collect their natural-language descriptions with dimensional information using Amazon Mechanical Turk. We positioned the CAD program (programming script for CAD design) as a bridge, facilitating the conversion of LLMs’ textual output into tangible CAD design objects. We focus on two critical capabilities: the generation of syntactically correct CAD programs (Cap1) and the accuracy of the parsed 3D shapes (Cap2) quantified by intersection over union. The results show that both GPT-4 and GPT-4V demonstrate potential in 3D CAD generation. Specifically, on average, GPT-4V outperforms when processing only text-based input, exceeding the results obtained using multimodal inputs, such as text with image, for Cap 1 and Cap 2. However, when examining category-specific results of mechanical components, while the same trend still holds for Cap 2, the prominence of multimodal inputs is increasingly evident for more complex geometries (e.g., springs and gears) in Cap 1. The potential of multimodal LLMs in enhancing 3D CAD generation is clear, but their application must be carefully calibrated to the complexity of the target CAD models to be generated.more » « less
-
Human-conducted rating tasks are resource-intensive and demand significant time and financial commitments. As Large Language Models (LLMs) like GPT emerge and exhibit prowess across various domains, their potential in automating such evaluation tasks becomes evident. In this research, we leveraged four prominent LLMs: GPT-4, GPT-3.5, Vicuna, and PaLM 2, to scrutinize their aptitude in evaluating teacher-authored mathematical explanations. We utilized a detailed rubric that encompassed accuracy, explanation clarity, the correctness of mathematical notation, and the efficacy of problem-solving strategies. During our investigation, we unexpectedly discerned the influence of HTML formatting on these evaluations. Notably, GPT-4 consistently favored explanations formatted with HTML, whereas the other models displayed mixed inclinations. When gauging Inter-Rater Reliability (IRR) among these models, only Vicuna and PaLM 2 demonstrated high IRR using the conventional Cohen’s Kappa metric for explanations formatted with HTML. Intriguingly, when a more relaxed version of the metric was applied, all model pairings showcased robust agreement. These revelations not only underscore the potential of LLMs in providing feedback on student-generated content but also illuminate new avenues, such as reinforcement learning, which can harness the consistent feedback from these models.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Conference Paper:
- The DOI is not currently available.
