Edited and reviewed by: Thomas Hartung, Johns Hopkins University, United States
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Since their inception in the 1980s, language models (LMs) have been around for more than four decades as a means for statistically modeling the properties observed from natural language (Rosenfeld,
It has only been recently that, as a consequence of the increase in the availability of text collections and in the access to improved computational resources, large language models (LLMs) have been introduced in the scientific community by revolutionizing the NLP field (Min et al.,
The surge of LLMs has been incremental since the late 2010s and has come in waves. Following a wave that introduced word embedding models such as word2vec (Mikolov et al.,
In recent years, LLMs have demonstrated to achieve state-of-the-art performance across many NLP tasks, having in turn become the
The great performance of LLMs has also inevitably provoked some fear in society that artificial intelligence tools may eventually take up many people's jobs (George et al.,
The success of LLMs is not without controversy, which is in turn shaping up ongoing research in NLP and opening up avenues for more research in improving these LLMs. The following are some of the key limitations of LLMs which need further exploration.
After the release of the first major LLM-based chatbot system that garnered mainstream popularity, OpenAI's ChatGPT, concerns emerged around the black box nature of the system. Indeed, there is no publicly available information on how ChatGPT was implemented as well as what data they used for training their model. From the perspective of NLP researchers, this raises serious concerns about the transparency and reproducibility of such model, not only because one does not know what is going on in the model, but also because it hinders reproducibility (Belz et al.,
To mitigate the impact, and increase our understanding, of black box models like ChatGPT, researchers have started investigating methods for reverse engineering those models, for example by trying to find out what data a model may have used for training (Shi et al.,
Luckily, however, there is a recent surge of open source models in the NLP scientific community, which have led to the release of models like Facebook's LLaMa 2 (Touvron et al.,
Data contamination occurs when “downstream test sets find their way into the pretrain corpus” (Magar and Schwartz,
The use of large-scale datasets for training LLMs also means that those datasets are very likely to contain biased or stereotyped information, which has been shown that LLMs amplify (Gallegos et al.,
Biases inherent in LLMs are at times exacerbated to even generate content that can be deemed offensive (Weidinger et al.,
Some systems, such as OpenAI's ChatGPT, acknowledge the risk of producing offensive content in their terms of service
“Our Services may provide incomplete, incorrect, or offensive Output that does not represent OpenAIs views. If Output references any third party products or services, it doesnt mean the third party endorses or is affiliated with OpenAI.”
LLMs can also capture sensitive information retrieved from its training data. While this information is encoded in embeddings which are not human readable, it has been found (Pan et al.,
Despite initial impressions that LLMs achieve an impressive performance, a closer look and investigation into model outputs shows that there is significant room for improvement. Evaluation of LLMs has in turn become a fertile area of research (Chang et al.,
Aware of the many shortcomings and inaccurate outputs of LLMs, companies responsible for the production and publication of major LLMs all have disclaimers about the limitations of their models. For example, ChatGPT owner OpenAI acknowledges that:
“Output may not always be accurate. You should not rely on Output from our Services as a sole source of truth or factual information, or as a substitute for professional advice.”
Google also warns
“Bard is an experimental technology and may sometimes give inaccurate or inappropriate information that doesnt represent Googles views.”
“Dont rely on Bards responses as medical, legal, financial, or other professional advice.”
Facebook also has a similar disclaimer
“Llama 2s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 2, developers should perform safety testing and tuning tailored to their specific applications of the model.”
Responses and outputs generated by LLMs often deviate from common sense, where for example a generated text can start discussing a particular topic, then shifting to another unrelated topic which is not intuitive, or even stating wrong facts. LLM hallucination has been defined as “the generation of content that deviates from the real facts, resulting in unfaithful outputs” (Maynez et al.,
The complexity of LLM models means that it is often very difficult to understand why it makes certain predictions or produces certain outputs. This also means that it is very difficult to provide explanations on model outputs to system users, which calls for more investigation into furthering the explainability of LLMs (Danilevsky et al.,
The introduction and surge in popularity of LLMs has impacted and reshaped NLP research. Much of the NLP research and methods slightly over a decade ago focused on the representation of words using bag-of-words and TF-IDF based methods and the use of machine learning algorithms such as Logistic Regression or Support Vector Machine classifiers. The increase in computational capacity to handle large-scale datasets and for more complex computing has led to the renaissance of deep learning models and in turn the emergence of LLMs. The latter have shown to achieve unprecedented performance across a range of downstream NLP tasks, but have also opened up numerous avenues for future research aiming to tackle the limitations and weaknesses of LLMs. Much of this research will need to deal with the better curation of the data fed to train LLMs, which in the current circumstances has shown to have severe risks in aspects such as fairness, privacy and harm.
AZ: Writing – original draft, Writing – review & editing.
The author(s) declare that no financial support was received for the research, authorship, and/or publication of this article.
The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
1
2
3