ChatGPT, a widely-recognized large language model (LLM), has recently gained substantial attention for its performance scaling, attributed to the billions of web-sourced natural language sentences used for training. Its underlying architecture, Transformer, has found applications across diverse fields, including video, audio signals, and robotic movement. %The crucial question this raises concerns the Transformer's generalization-in-learning (GIL) capacity. However, this raises a crucial question about Transformer's generalization in learning (GIL) capacity. Is ChatGPT's success chiefly due to the vast dataset used for training, or is there more to the story? To investigate this, we compared Transformer's GIL capabilities with those of a traditional Recurrent Neural Network (RNN) in tasks involving attractor dynamics learning. For performance evaluation, the Dynamic Time Warping (DTW) method has been employed. Our simulation results suggest that under conditions of limited data availability, Transformer's GIL abilities are markedly inferior to those of RNN.

The rapid advancement of Large Language Models (LLMs) has sparked intense debate regarding their ability to perceive and interpret complex socio-political landscapes. In this study, we undertake an exploration of decisionmaking processes and inherent biases within Figure 1: (Undesired) Effect of Bias Treatment on Decision LLMs, exemplified by ChatGPT, specifically Process: The figure depicts how the LLM's perception contextualizing our analysis within political debates. of value A is considered during the decision We aim not to critique or validate LLMs' process while judging B and C through f(C|A) and values, but rather to discern how they interpret f(B|A). When treating the biased association of value and adjudicate "good arguments." By applying A with C (f(C|A)) by naively fine-tuning the model to Activity Dependency Networks (ADNs), align with this value of interest, other value associations we extract the LLMs' implicit criteria for such (f(B|A)), that are not actively considered. They may assessments and illustrate how normative values be changed indiscriminately, regardless of whether they influence these perceptions. We discuss were already aligned. These associations are currently the consequences of our findings for human-AI neither observable nor predictable yet changes in them alignment and bias mitigation.

Large language models (LLMs) have demonstrated remarkable open-domain capabilities. Traditionally, LLMs tailored for a domain are trained from scratch to excel at handling domain-specific tasks. In this work, we explore an alternative strategy of continual pre-training as a means to develop domain-specific LLMs. We introduce FinPythia-6.9B, developed through domain-adaptive continual pre-training on the financial domain. Continual pre-trained FinPythia showcases consistent improvements on financial tasks over the original foundational model. We further explore simple but effective data selection strategies for continual pre-training. Our data selection strategies outperforms vanilla continual pre-training's performance with just 10% of corpus size and cost, without any degradation on open-domain standard tasks. Our work proposes an alternative solution to building domain-specific LLMs from scratch in a cost-effective manner.

Applications of Transformer models are exploding, e.g., ChatGPT [1]. Security is critical to Transformer-based services, which determines whether applications can be scaled to privacy-sensitive areas like cloud copilot for proprietary code and documents [2]. Existing work [3, 4] studied this problem under the classic secure multi-party computing framework. Using encryption and decryption methods requires approximation of complex nonlinear layers and introduces heavy computational overhead. In this work, we propose a three-party protocol using permutation to protect both model parameters and user data without any approximation of Transformer models.

Reciprocal questioning is essential for effective teaching and learning, fostering active engagement and deeper understanding through collaborative interactions, especially in large classrooms. Can large language model (LLM), such as OpenAI's GPT (Generative Pre-trained Transformer) series, assist in this? This paper investigates a pedagogical approach of classroom flipping based on flipped interaction in LLMs. Flipped interaction involves using language models to prioritize generating questions instead of answers to prompts. We demonstrate how traditional classroom flipping techniques, including Peer Instruction and Just-in-Time Teaching (JiTT), can be enhanced through flipped interaction techniques, creating student-centric questions for hybrid teaching. In particular, we propose a workflow to integrate prompt engineering with clicker and JiTT quizzes by a poll-prompt-quiz routine and a quiz-prompt-discuss routine to empower students to self-regulate their learning capacity and enable teachers to swiftly personalize training pathways. We develop an LLM-driven chatbot software that digitizes various elements of classroom flipping and facilitates the assessment of students using these routines to deliver peer-generated questions. We have applied our LLM-driven chatbot software for teaching both undergraduate and graduate students from 2020 to 2022, effectively useful for bridging the gap between teachers and students in remote teaching during the COVID-19 pandemic years. In particular, LLM-driven classroom flipping can be particularly beneficial in large class settings to optimize teaching pace and enable engaging classroom experiences.

Artificial general intelligence (AGI) does not yet exist, but given the pace of technological development in artificial intelligence, it is projected to reach human-level intelligence within roughly the next two decades. After that, many experts expect it to far surpass human intelligence and to do so rapidly. The prospect of superintelligent AGI poses an existential risk to humans because there is no reliable method for ensuring that AGI goals stay aligned with human goals. Drawing on publicly available forecaster and opinion data, the author examines how experts and non-experts perceive risk from AGI. The findings indicate that the perceived risk of a world catastrophe or extinction from AGI is greater than for other existential risks. The increase in perceived risk over the last year is also steeper for AGI than for other existential threats (e.g., nuclear war or human-caused climate change). That AGI is a pressing existential risk is something on which experts and non-experts agree, but the basis for such agreement currently remains obscure.

The complementary potential of Large Language Models (LLM) assumes off-the-shelf LLMs have heterogeneous expertise in a wide range of domains and tasks so that an ensemble of LLMs can achieve consistently better performance. Existing ensemble methods for LLMs mainly focus on reward model ranking of outputs, leading to significant computation overhead. To combat this issue, we revisit the complementary potential of LLMs and further elaborate it by mining latent expertise with off-the-shelf reward models. We propose Zooter, a reward-guided routing method distilling rewards on training queries to train a routing function, which can precisely distribute each query to the LLM with expertise about it. We also integrate a tag-based label enhancement to mitigate noise from uncertainty when using rewards as silver supervision. Zooter shows computation efficiency in inference as it introduces only a minor computation overhead of a routing function compared with reward model ranking methods. We evaluate Zooter on a comprehensive benchmark collection with 26 subsets on different domains and tasks. Zooter outperforms the best single model on average and ranks first on 44% of tasks, even surpassing multiple reward model ranking methods.

Diabetic eye disease is a major cause of blindness worldwide. The ability to monitor relevant clinical trajectories and detect lapses in care is critical to managing the disease and preventing blindness. Alas, much of the information necessary to support these goals is found only in the free text of the electronic medical record. To fill this information gap, we introduce a system for extracting evidence from clinical text of 19 clinical concepts related to diabetic eye disease and inferring relevant attributes for each. In developing this ophthalmology phenotyping system, we are also afforded a unique opportunity to evaluate the effectiveness of clinical language models at adapting to new clinical domains. Across multiple training paradigms, we find that BERT language models pretrained on out-of-distribution clinical data offer no significant improvement over BERT language models pretrained on non-clinical data for our domain. Our study tempers recent claims that language models pretrained on clinical data are necessary for clinical NLP tasks and highlights the importance of not treating clinical language data as a single homogeneous domain.

Reinforcement Learning from Human Feedback (RLHF) is a vital strategy for enhancing model safety in language models. However, annotating preference data for RLHF is a resource-intensive and creativity-demanding process, while automatic generation methods face limitations in data diversity and quality. In response, we present Safer-Instruct, a novel pipeline for semi-automatically constructing large-scale preference datasets. Our approach leverages reversed instruction tuning, instruction induction, and expert model evaluation to efficiently generate high-quality preference data without human annotators. We evaluate Safer-Instruct using LLaMA for instruction induction and GPT-4 as an expert model, generating approximately 10K preference samples. Finetuning an Alpaca model on this dataset demonstrates improved harmlessness while maintaining competitive performance on conversation and downstream tasks. Safer-Instruct addresses the challenges in preference data acquisition, advancing the development of safer and more responsible AI systems. Our code and data are available at https://github.com/uscnlp-lime/safer-instruct

Multilingual pre-trained language models are incredibly effective at Question Answering (QA), a core task in Natural Language Understanding, achieving high accuracies on several multilingual benchmarks. However, little is known about how well they are calibrated. In this paper, we study the calibration properties of several pre-trained multilingual large language models (LLMs) on a variety of question-answering tasks. We perform extensive experiments, spanning both extractive and generative QA model designs and diverse languages, spanning both high-resource and low-resource ones. We study different dimensions of calibration in in-distribution, out-of-distribution, and cross-lingual transfer settings, and investigate strategies to improve it, including post-hoc methods and regularized fine-tuning. We demonstrate automatically translated data augmentation as a highly effective technique to improve model calibration. We also conduct a number of ablation experiments to study the effect of model size on calibration and how multilingual models compare with their monolingual counterparts for diverse tasks and languages.