As large language models (LLMs) perform more difficult tasks, it becomes harder to verify the correctness and safety of their behavior. One approach to help with this issue is to prompt LLMs to externalize their reasoning, e.g., by having them generate step-by-step reasoning as they answer a question (Chain-of-Thought; CoT). The reasoning may enable us to check the process that models use to perform tasks. However, this approach relies on the stated reasoning faithfully reflecting the model's actual reasoning, which is not always the case. To improve over the faithfulness of CoT reasoning, we have models generate reasoning by decomposing questions into subquestions. Decomposition-based methods achieve strong performance on question-answering tasks, sometimes approaching that of CoT while improving the faithfulness of the model's stated reasoning on several recently-proposed metrics. By forcing the model to answer simpler subquestions in separate contexts, we greatly increase the faithfulness of model-generated reasoning over CoT, while still achieving some of the performance gains of CoT. Our results show it is possible to improve the faithfulness of model-generated reasoning; continued improvements may lead to reasoning that enables us to verify the correctness and safety of LLM behavior.

Bias evaluation benchmarks and dataset and model documentation have emerged as central processes for assessing the biases and harms of artificial intelligence (AI) systems. However, these auditing processes have been criticized for their failure to integrate the knowledge of marginalized communities and consider the power dynamics between auditors and the communities. Consequently, modes of bias evaluation have been proposed that engage impacted communities in identifying and assessing the harms of AI systems (e.g., bias bounties). Even so, asking what marginalized communities want from such auditing processes has been neglected. In this paper, we ask queer communities for their positions on, and desires from, auditing processes. To this end, we organized a participatory workshop to critique and redesign bias bounties from queer perspectives. We found that when given space, the scope of feedback from workshop participants goes far beyond what bias bounties afford, with participants questioning the ownership, incentives, and efficacy of bounties. We conclude by advocating for community ownership of bounties and complementing bounties with participatory processes (e.g., co-creation).

Current transformer language models (LM) are large-scale models with billions of parameters. They have been shown to provide high performances on a variety of tasks but are also prone to shortcut learning and bias. Addressing such incorrect model behavior via parameter adjustments is very costly. This is particularly problematic for updating dynamic concepts, such as moral values, which vary culturally or interpersonally. In this work, we question the current common practice of storing all information in the model parameters and propose the Revision Transformer (RiT) to facilitate easy model updating. The specific combination of a large-scale pre-trained LM that inherently but also diffusely encodes world knowledge with a clear-structured revision engine makes it possible to update the model's knowledge with little effort and the help of user interaction. We exemplify RiT on a moral dataset and simulate user feedback demonstrating strong performance in model revision even with small data. This way, users can easily design a model regarding their preferences, paving the way for more transparent AI models.

The AI powering ChatGPT may provide completely different answers to the same mathematical problems over time. Those findings from recent experiments have fuelled an ongoing debate about whether the AI chatbot's performance is getting worse – and have spurred the firm behind it, OpenAI, to reassure customers that applications built on ChatGPT will not continually break. "The takeaway message is that the behaviour of the'same' large language model can change substantially," says Lingjiao Chen at Stanford University in California.

With the boom in popularity of artificial intelligence (AI), attention has quickly turned to the impact such innovation could have on the jobs market. There are fears that thousands of human roles may soon disappear because of huge advances in automation, with it emerging only last month that the UK Government privately thinks a'substantial number' of civil service jobs will soon be obsolete. Not to mention there is also a growing Silicon Valley civil war about whether rapidly evolving AI technology is a good thing or a bad thing. Well, research suggests that air traffic controllers, midwives, librarians and those with a career in sales have little to worry about, but if you work behind a bar, as a window cleaner or in customer service, the news isn't quite so positive. Nor is it for waiters and waitresses, who at 72 per cent are at the highest risk of having their roles carried out by a robot, according to digital media company DailyAI.com.

On a Monday morning in April, Sam Altman sat inside OpenAI's San Francisco headquarters, telling me about a dangerous artificial intelligence that his company had built but would never release. His employees, he later said, often lose sleep worrying about the AIs they might one day release without fully appreciating their dangers. With his heel perched on the edge of his swivel chair, he looked relaxed. The powerful AI that his company had released in November had captured the world's imagination like nothing in tech's recent history. There was grousing in some quarters about the things ChatGPT could not yet do well, and in others about the future it may portend, but Altman wasn't sweating it; this was, for him, a moment of triumph. Check out more from this issue and find your next story to read. In small doses, Altman's large blue eyes emit a beam of earnest intellectual attention, and he seems to understand that, in large doses, their intensity might unsettle. In this case, he was ...

We present a novel method for mining opinions from text collections using generative language models trained on data collected from different populations. We describe the basic definitions, methodology and a generic algorithm for opinion insight mining. We demonstrate the performance of our method in an experiment where a pre-trained generative model is fine-tuned using specifically tailored content with unnatural and fully annotated opinions. We show that our approach can learn and transfer the opinions to the semantic classes while maintaining the proportion of polarisation. Finally, we demonstrate the usage of an insight mining system to scale up the discovery of opinion insights from a real text corpus.

In this paper, we present a novel framework that combines large language models (LLMs), digital twins and industrial automation system to enable intelligent planning and control of production processes. We retrofit the automation system for a modular production facility and create executable control interfaces of fine-granular functionalities and coarse-granular skills. Low-level functionalities are executed by automation components, and high-level skills are performed by automation modules. Subsequently, a digital twin system is developed, registering these interfaces and containing additional descriptive information about the production system. Based on the retrofitted automation system and the created digital twins, LLM-agents are designed to interpret descriptive information in the digital twins and control the physical system through service interfaces. These LLM-agents serve as intelligent agents on different levels within an automation system, enabling autonomous planning and control of flexible production. Given a task instruction as input, the LLM-agents orchestrate a sequence of atomic functionalities and skills to accomplish the task. We demonstrate how our implemented prototype can handle un-predefined tasks, plan a production process, and execute the operations. This research highlights the potential of integrating LLMs into industrial automation systems in the context of smart factory for more agile, flexible, and adaptive production processes, while it also underscores the critical insights and limitations for future work. Demos at: https://github.com/YuchenXia/GPT4IndustrialAutomation

Large language models such as ChatGPT-3.5 and GPT-4.0 are ubiquitous and dominate the current discourse. Their transformative capabilities have led to a paradigm shift in how we interact with and utilize (text-based) information. Each day, new possibilities to leverage the capabilities of these models emerge. This paper presents findings on the performance of different large language models in a university of applied sciences' undergraduate computer science degree program. Our primary objective is to assess the effectiveness of these models within the curriculum by employing them as educational aids. By prompting the models with lecture material, exercise tasks, and past exams, we aim to evaluate their proficiency across different computer science domains. We showcase the strong performance of current large language models while highlighting limitations and constraints within the context of such a degree program. We found that ChatGPT-3.5 averaged 79.9% of the total score in 10 tested modules, BingAI achieved 68.4%, and LLaMa, in the 65 billion parameter variant, 20%. Despite these convincing results, even GPT-4.0 would not pass the degree program - due to limitations in mathematical calculations.

Large language models have made significant progress in the past few years. However, they are either generic {\it or} field specific, splitting the community into different groups. In this paper, we unify these large language models into a larger map, where the generic {\it and} specific models are linked together and can improve each other, based on the user personal input and information from the internet. The idea of linking several large language models together is inspired by the functionality of human brain. The specific regions on the brain cortex are specific for certain low level functionality. And these regions can jointly work together to achieve more complex high level functionality. Such behavior on human brain cortex sheds the light to design the multilevel large language models that contain global level, field level and user level models. The user level models run on local machines to achieve efficient response and protect the user's privacy. Such multilevel models reduce some redundancy and perform better than the single level models. The proposed multilevel idea can be applied in various applications, such as natural language processing, computer vision tasks, professional assistant, business and healthcare.