The widespread adoption of Large Language Models (LLMs) has become commonplace, particularly with the emergence of open-source models. More importantly, smaller models are well-suited for integration into consumer devices and are frequently employed either as standalone solutions or as subroutines in various AI tasks. Despite their ubiquitous use, there is no systematic analysis of their specific capabilities and limitations. In this study, we tackle one of the most widely used tasks - answering Multiple Choice Question (MCQ). We analyze 26 small open-source models and find that 65% of the models do not understand the task, only 4 models properly select an answer from the given choices, and only 5 of these models are choice order independent. These results are rather alarming given the extensive use of MCQ tests with these models. We recommend exercising caution and testing task understanding before using MCQ to evaluate LLMs in any field whatsoever.

\label{sec:abstract} Large Language Models (LLMs) have shown promise in assisting scientific discovery. However, such applications are currently limited by LLMs' deficiencies in understanding intricate scientific concepts, deriving symbolic equations, and solving advanced numerical calculations. To bridge these gaps, we introduce SciGLM, a suite of scientific language models able to conduct college-level scientific reasoning. Central to our approach is a novel self-reflective instruction annotation framework to address the data scarcity challenge in the science domain. This framework leverages existing LLMs to generate step-by-step reasoning for unlabelled scientific questions, followed by a process of self-reflective critic-and-revise. Applying this framework, we curated SciInstruct, a diverse and high-quality dataset encompassing mathematics, physics, chemistry, and formal proofs. We fine-tuned the ChatGLM family of language models with SciInstruct, enhancing their capabilities in scientific and mathematical reasoning. Remarkably, SciGLM consistently improves both the base model (ChatGLM3-6B-Base) and larger-scale models (12B and 32B), without sacrificing the language understanding capabilities of the base model. This makes SciGLM a suitable foundational model to facilitate diverse scientific discovery tasks. For the benefit of the wider research community, we release SciInstruct, SciGLM, alongside a self-reflective framework and fine-tuning code at \url{https://github.com/THUDM/SciGLM}.

We propose an algorithm for optimizing the parameters of single hidden layer neural networks. Specifically, we derive a blockwise difference-of-convex (DC) functions representation of the objective function. Based on the latter, we propose a block coordinate descent (BCD) approach that we combine with a tailored difference-of-convex functions algorithm (DCA). We prove global convergence of the proposed algorithm. Furthermore, we mathematically analyze the convergence rate of parameters and the convergence rate in value (i.e., the training loss). We give conditions under which our algorithm converges linearly or even faster depending on the local shape of the loss function. We confirm our theoretical derivations numerically and compare our algorithm against state-of-the-art gradient-based solvers in terms of both training loss and test loss.

Incels are an extremist online community of men who believe in an ideology rooted in misogyny, racism, the glorification of violence, and dehumanization. In their online forums, they use an extensive, evolving cryptolect - a set of ingroup terms that have meaning within the group, reflect the ideology, demonstrate membership in the community, and are difficult for outsiders to understand. This paper presents a lexicon with terms and definitions for common incel root words, prefixes, and affixes. The lexicon is text-based for use in automated analysis and is derived via a Qualitative Content Analysis of the most frequent incel words, their structure, and their meaning on five of the most active incel communities from 2016 to 2023.

Human dexterity is an invaluable capability for precise manipulation of objects in complex tasks. The capability of robots to similarly grasp and perform in-hand manipulation of objects is critical for their use in the ever changing human environment, and for their ability to replace manpower. In recent decades, significant effort has been put in order to enable in-hand manipulation capabilities to robotic systems. Initial robotic manipulators followed carefully programmed paths, while later attempts provided a solution based on analytical modeling of motion and contact. However, these have failed to provide practical solutions due to inability to cope with complex environments and uncertainties. Therefore, the effort has shifted to learning-based approaches where data is collected from the real world or through a simulation, during repeated attempts to complete various tasks. The vast majority of learning approaches focused on learning data-based models that describe the system to some extent or Reinforcement Learning (RL). RL, in particular, has seen growing interest due to the remarkable ability to generate solutions to problems with minimal human guidance. In this survey paper, we track the developments of learning approaches for in-hand manipulations and, explore the challenges and opportunities. This survey is designed both as an introduction for novices in the field with a glossary of terms as well as a guide of novel advances for advanced practitioners.

The advent of Large Language Models (LLMs) represents a notable breakthrough in Natural Language Processing (NLP), contributing to substantial progress in both text comprehension and generation. However, amidst these advancements, it is noteworthy that LLMs often face a limitation in terms of context length extrapolation. Understanding and extending the context length for LLMs is crucial in enhancing their performance across various NLP applications. In this survey paper, we delve into the multifaceted aspects of exploring why it is essential, and the potential transformations that superior techniques could bring to NLP applications. We study the inherent challenges associated with extending context length and present an organized overview of the existing strategies employed by researchers. Additionally, we discuss the intricacies of evaluating context extension techniques and highlight the open challenges that researchers face in this domain. Furthermore, we explore whether there is a consensus within the research community regarding evaluation standards and identify areas where further agreement is needed. This comprehensive survey aims to serve as a valuable resource for researchers, guiding them through the nuances of context length extension techniques and fostering discussions on future advancements in this evolving field.

Large language models show compelling performance on reasoning tasks but they tend to perform much worse in languages other than English. This is unsurprising given that their training data largely consists of English text and instructions. A typical solution is to translate instruction data into all languages of interest, and then train on the resulting multilingual data, which is called translate-training. This approach not only incurs high cost, but also results in poorly translated data due to the non-standard formatting of chain-of-thought and mathematical reasoning instructions. In this paper, we explore the benefits of question alignment, where we train the model to translate reasoning questions into English by finetuning on X-English question data. In this way we perform targetted, in-domain language alignment which makes best use of English instruction data to unlock the LLMs' multilingual reasoning abilities. Experimental results on LLaMA2-13B show that question alignment leads to consistent improvements over the translate-training approach: an average improvement of 11.3\% and 16.1\% accuracy across ten languages on the MGSM and MSVAMP maths reasoning benchmarks (The project will be available at: https://github.com/NJUNLP/QAlign).

Wikidata has grown to a knowledge graph with an impressive size. To date, it contains more than 17 billion triples collecting information about people, places, films, stars, publications, proteins, and many more. On the other side, most of the information on the Web is not published in highly structured data repositories like Wikidata, but rather as unstructured and semi-structured content, more concretely in HTML pages containing text and tables. Finding, monitoring, and organizing this data in a knowledge graph is requiring considerable work from human editors. The volume and complexity of the data make this task difficult and time-consuming. In this work, we present a framework that is able to identify and extract new facts that are published under multiple Web domains so that they can be proposed for validation by Wikidata editors. The framework is relying on question-answering technologies. We take inspiration from ideas that are used to extract facts from textual collections and adapt them to extract facts from Web pages. For achieving this, we demonstrate that language models can be adapted to extract facts not only from textual collections but also from Web pages. By exploiting the information already contained in Wikidata the proposed framework can be trained without the need for any additional learning signals and can extract new facts for a wide range of properties and domains. Following this path, Wikidata can be used as a seed to extract facts on the Web. Our experiments show that we can achieve a mean performance of 84.07 at F1-score. Moreover, our estimations show that we can potentially extract millions of facts that can be proposed for human validation. The goal is to help editors in their daily tasks and contribute to the completion of the Wikidata knowledge graph.

AI agents based on multimodal large language models (LLMs) are expected to revolutionize human-computer interaction and offer more personalized assistant services across various domains like healthcare, education, manufacturing, and entertainment. Deploying LLM agents in 6G networks enables users to access previously expensive AI assistant services via mobile devices democratically, thereby reducing interaction latency and better preserving user privacy. Nevertheless, the limited capacity of mobile devices constrains the effectiveness of deploying and executing local LLMs, which necessitates offloading complex tasks to global LLMs running on edge servers during long-horizon interactions. In this article, we propose a split learning system for LLM agents in 6G networks leveraging the collaboration between mobile devices and edge servers, where multiple LLMs with different roles are distributed across mobile devices and edge servers to perform user-agent interactive tasks collaboratively. In the proposed system, LLM agents are split into perception, grounding, and alignment modules, facilitating inter-module communications to meet extended user requirements on 6G network functions, including integrated sensing and communication, digital twins, and task-oriented communications. Furthermore, we introduce a novel model caching algorithm for LLMs within the proposed system to improve model utilization in context, thus reducing network costs of the collaborative mobile and edge LLM agents.

Thanks to recent advances in generative AI, we are able to prompt large language models (LLMs) to produce texts which are fluent and grammatical. In addition, it has been shown that we can elicit attempts at grammatical error correction (GEC) from LLMs when prompted with ungrammatical input sentences. We evaluate how well LLMs can perform at GEC by measuring their performance on established benchmark datasets. We go beyond previous studies, which only examined GPT* models on a selection of English GEC datasets, by evaluating seven open-source and three commercial LLMs on four established GEC benchmarks. We investigate model performance and report results against individual error types. Our results indicate that LLMs do not always outperform supervised English GEC models except in specific contexts -- namely commercial LLMs on benchmarks annotated with fluency corrections as opposed to minimal edits. We find that several open-source models outperform commercial ones on minimal edit benchmarks, and that in some settings zero-shot prompting is just as competitive as few-shot prompting.