Foundation and large language models: fundamentals, challenges, opportunities, and social impacts

Myers, Devon; Mohawesh, Rami; Chellaboina, Venkata Ishwarya; Sathvik, Anantha Lakshmi; Venkatesh, Praveen; Ho, Yi-Hui; Henshaw, Hanna; Alhawawreh, Muna; Berdik, David; Jararweh, Yaser

doi:10.1007/s10586-023-04203-7

Cited by 16 publications

(8 citation statements)

References 141 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This suggests that language models serve as invaluable allies in the pursuit of writing excellence, offering indispensable support in identifying and rectifying errors and inconsistencies. This result aligns with that of the prior investigation (Myers et al, 2023) which found that Large language models can be used to identify (possible) grammatical problems at the semantic level and recommend appropriate and customized remediation techniques. The simplest assistance can be provided at the syntactic level, which involves finding and fixing errors.…”

Section: "Language Models Such As Grammarly Helped Me Meticulously Re...supporting

confidence: 89%

Exploring the Impact of Language Models on Undergraduate Thesis Writing Skills among Higher Education Students

Ilma,

Riza Ady Sampurna

2024

PJIER

View full text Add to dashboard Cite

This article explores the impact of language models on undergraduate thesis writing skills among higher education students. With advancements in technology, language models have become integral to the digital writing landscape. This study examined the influence of language models on students' confidence and motivation in academic writing. The research employs a qualitative approach, conducting semi-structured interviews with five students from an English Language Education program who utilized AI tools in their thesis writing process. Despite acknowledging the advantages, participants underscored challenges encompassing overreliance, inaccuracies, and disruptions in the natural flow of their writing. The impact of these tools on participants' confidence and motivation in academic writing manifested diversely, introducing varying degrees of perceived artificiality in writing proficiency. There were already numerous researches regarding the pretinence of language models usage on students’ writing aptitude. However, there remains a gap in understanding their specific influence on undergraduate thesis writing. Thematic analysis of the interview data reveals insights into the influence of language models, on the writing proficiency of higher education students engaged in the process of crafting undergraduate theses. Employing a methodology encompassing interviews, participants reported noteworthy enhancements in the coherence, clarity, grammar, syntax, and vocabulary of their theses facilitated by the utilization of these language tools. Overall, the incorporation of language models in the writing process emerged as a catalyst for heightened efficiency, skills development, and enhanced clarity in undergraduate theses, affirming the constructive role of AI technology in educational settings. Individualized experiences with language models exhibited variance contingent upon personal preferences and attitudes towards the writing process.

show abstract

Section: "Language Models Such As Grammarly Helped Me Meticulously Re...supporting

confidence: 89%

Exploring the Impact of Language Models on Undergraduate Thesis Writing Skills among Higher Education Students

Ilma,

Riza Ady Sampurna

2024

PJIER

View full text Add to dashboard Cite

show abstract

“…Another notable development was the integration of machine learning techniques that focused on anomaly detection within the generated text, flagging content that deviated significantly from established facts or logical coherence [39], [40], [41]. Despite the promise of these methodologies, they often required extensive computational resources and faced challenges in scaling up for broader applications [42]. This limitation highlighted the need for more efficient and scalable solutions that could maintain high standards of accuracy without compromising on the generative efficiency of LLMs, paving the way for innovative approaches like the integration of Cross-Referential Validation modules in models such as GPT-Neo-CRV.…”

Section: Existing Methods For Enhancing Llm Reliabilitymentioning

confidence: 99%

GPT-Neo-CRV: Elevating Information Accuracy in GPT-Neo with Cross-Referential Validation

Xiong,

Zheng

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Studies within this theme have consistently demonstrated that LLMs can inherit and amplify biases present in their training data, affecting their language generation and decisionmaking processes [4], [8], [23]- [25]. A common finding across several studies is that biases related to gender, race, and ethnicity are particularly prevalent, often leading to stereotypical or prejudicial outputs [12], [20], [21], [26], [27].…”

Section: A Identification and Analysis Of Bias In Llmsmentioning

confidence: 99%

A Comparative Analysis to Evaluate Bias and Fairness Across Large Language Models with Benchmarks

陳文意,

黃兆明

2024

Preprint

View full text Add to dashboard Cite

This study performs a comprehensive evaluation of bias and fairness within Large Language Models (LLMs), including ChatGPT-4, Google Gemini, and Llama 2, utilizing the Google BIG-Bench benchmark. Our analysis reveals varied levels of biases across models, with disparities particularly notable in dimensions such as gender, race, and ethnicity. The Google BIG-Bench benchmark proved instrumental in identifying these biases, though its effectiveness is tempered by challenges in capturing the sophisticated manifestations of bias that emerge in real-world contexts. Comparative performance analysis indicates that while each model exhibits strengths in certain areas, no single model uniformly excels across all fairness and bias metrics. The study underscores the intricate balance between model performance, fairness, and efficiency, highlighting the necessity for ongoing research and development in AI ethics to mitigate bias effectively. Insights from this research advocate for a multifaceted approach to AI development, integrating ethical considerations at every stage to ensure the equitable advancement of technology. The findings prompt a call for continued innovation in model training and benchmarking methodologies, aiming to enhance the fairness and inclusivity of future LLMs.

show abstract

Foundation and large language models: fundamentals, challenges, opportunities, and social impacts

Cited by 16 publications

References 141 publications

Exploring the Impact of Language Models on Undergraduate Thesis Writing Skills among Higher Education Students

Exploring the Impact of Language Models on Undergraduate Thesis Writing Skills among Higher Education Students

GPT-Neo-CRV: Elevating Information Accuracy in GPT-Neo with Cross-Referential Validation

A Comparative Analysis to Evaluate Bias and Fairness Across Large Language Models with Benchmarks

Contact Info

Product

Resources

About