To better understand existing LLMs, we may examine the human mental (cognitive/psychological) architecture, and its components and structures. Based on psychological, philosophical, and cognitive science literatures, it is argued that, within the human mental architecture, existing LLMs correspond well with implicit mental processes (intuition, instinct, and so on). However, beyond such implicit processes, explicit processes (with better symbolic capabilities) are also present within the human mental architecture, judging from psychological, philosophical, and cognitive science literatures. Various theoretical and empirical issues and questions in this regard are explored. Furthermore, it is argued that existing dual-process computational cognitive architectures (models of the human cognitive/psychological architecture) provide usable frameworks for fundamentally enhancing LLMs by introducing dual processes (both implicit and explicit) and, in the meantime, can also be enhanced by LLMs. The results are synergistic combinations (in several different senses simultaneously).

Recent advancements in Direct Preference Optimization (DPO) have significantly enhanced the alignment of Large Language Models (LLMs) with human preferences, owing to its simplicity and effectiveness. However, existing methods typically optimize a scalar score or ranking reward, thereby overlooking the multi-dimensional nature of human preferences. In this work, we propose to extend the preference of DPO to two dimensions: segments and aspects. We first introduce a 2D supervision dataset called HelpSteer-2D. For the segment dimension, we divide the response into sentences and assign scores to each segment. For the aspect dimension, we meticulously design several criteria covering the response quality rubrics. With the 2-dimensional signals as feedback, we develop a 2D-DPO framework, decomposing the overall objective into multi-segment and multi-aspect objectives. Extensive experiments on popular benchmarks demonstrate that 2D-DPO performs better than methods that optimize for scalar or 1-dimensional preferences.

The application of artificial intelligence (AI) in civil engineering presents a transformative approach to enhancing design quality and safety. This paper investigates the potential of the advanced LLM GPT4 Turbo vision model in detecting architectural flaws during the design phase, with a specific focus on identifying missing doors and windows. The study evaluates the model's performance through metrics such as precision, recall, and F1 score, demonstrating AI's effectiveness in accurately detecting flaws compared to human-verified data. Additionally, the research explores AI's broader capabilities, including identifying load-bearing issues, material weaknesses, and ensuring compliance with building codes. The findings highlight how AI can significantly improve design accuracy, reduce costly revisions, and support sustainable practices, ultimately revolutionizing the civil engineering field by ensuring safer, more efficient, and aesthetically optimized structures.

Visual Question-Answering, a technology that generates textual responses from an image and natural language question, has progressed significantly. Notably, it can aid in tracking and inquiring about daily activities, crucial in healthcare monitoring, especially for elderly patients or those with memory disabilities. However, video poses privacy concerns and has a limited field of view. This paper presents Sensor2Text, a model proficient in tracking daily activities and engaging in conversations using wearable sensors. The approach outlined here tackles several challenges, including low information density in wearable sensor data, insufficiency of single wearable sensors in human activities recognition, and model's limited capacity for Question-Answering and interactive conversations. To resolve these obstacles, transfer learning and student-teacher networks are utilized to leverage knowledge from visual-language models. Additionally, an encoder-decoder neural network model is devised to jointly process language and sensor data for conversational purposes. Furthermore, Large Language Models are also utilized to enable interactive capabilities. The model showcases the ability to identify human activities and engage in Q\&A dialogues using various wearable sensor modalities. It performs comparably to or better than existing visual-language models in both captioning and conversational tasks. To our knowledge, this represents the first model capable of conversing about wearable sensor data, offering an innovative approach to daily activity tracking that addresses privacy and field-of-view limitations associated with current vision-based solutions.

Large language models have evolved to process multiple modalities beyond text, such as images and audio, which motivates us to explore how to effectively leverage them for graph machine learning tasks. The key question, therefore, is how to transform graphs into linear sequences of tokens, a process we term graph linearization, so that LLMs can handle graphs naturally. We consider that graphs should be linearized meaningfully to reflect certain properties of natural language text, such as local dependency and global alignment, in order to ease contemporary LLMs, trained on trillions of textual tokens, better understand graphs. To achieve this, we developed several graph linearization methods based on graph centrality, degeneracy, and node relabeling schemes. We then investigated their effect on LLM performance in graph reasoning tasks. Experimental results on synthetic graphs demonstrate the effectiveness of our methods compared to random linearization baselines. Our work introduces novel graph representations suitable for LLMs, contributing to the potential integration of graph machine learning with the trend of multi-modal processing using a unified transformer model.

Integrating text and numbers effectively is a crucial step towards enhancing Large Language Models (LLMs) capabilities in assisting in scientific tasks. While most current approaches rely on discrete tokenization of numbers, for instance, conversion to scientific notation or base 10-decomposition, a recent approach proposed a continuous numerical encoding as an inductive bias. In this paper, we build upon this approach by introducing more expressive numerical embeddings. Our method addresses key shortcomings, including the elimination of numerical artefacts and the ability to handle a wide range of magnitudes without clipping. Our work presents two key contributions. First, we employ an MLP to assign distinct directions in the embedding space to different numbers. Our second contribution is the introduction of a routing layer that differentiates between numerical and text embeddings. We hypothesise that this combined approach enables the model to distinguish between text and number distributions while maintaining its capacity for arithmetic operations. Using only a 45 M parameter encoder-decoder architecture our method achieves a $R^2$=0.9988 over a wide range of magnitude ($10^{-3},10^{8}$). In addition, we empirically observe a reduction of the numerical artefacts and biases observed compared to the baselines.

Large language models (LLMs) can store a significant amount of factual knowledge in their parameters. However, their parametric knowledge may conflict with the information provided in the context -- this phenomenon, known as \emph{context-memory knowledge conflicts}, can lead to undesirable model behaviour, such as reliance on outdated or incorrect information. Analysing the internal activations of LLMs, we find that they can internally register the signals of knowledge conflict at mid-layers. Such signals allow us to detect whether a knowledge conflict occurs and use \emph{inference-time} intervention strategies to resolve it. In this work, we propose \textsc{SpARE}, a \emph{training-free} representation engineering method that uses pre-trained sparse auto-encoders (SAEs) to control the knowledge selection behaviour of LLMs. \textsc{SpARE} identifies the functional features that control the knowledge selection behaviours and applies them to edit the internal activations of LLMs at inference time. Our experimental results show that \textsc{SpARE} can effectively control the usage of either knowledge source to resolve knowledge conflict in open-domain question-answering tasks, surpassing existing representation engineering methods ($+10\%$) as well as contrastive decoding methods ($+15\%$).

The processes underlying human cognition are often divided into System 1, which involves fast, intuitive thinking, and System 2, which involves slow, deliberate reasoning. Previously, large language models were criticized for lacking the deeper, more analytical capabilities of System 2. In September 2024, OpenAI introduced the o1 model series, designed to handle System 2-like reasoning. While OpenAI's benchmarks are promising, independent validation is still needed. In this study, we tested the o1-preview model twice on the Dutch 'Mathematics B' final exam. It scored a near-perfect 76 and 74 out of 76 points. For context, only 24 out of 16,414 students in the Netherlands achieved a perfect score. By comparison, the GPT-4o model scored 66 and 62 out of 76, well above the Dutch students' average of 40.63 points. Neither model had access to the exam figures. Since there was a risk of model contami-nation (i.e., the knowledge cutoff for o1-preview and GPT-4o was after the exam was published online), we repeated the procedure with a new Mathematics B exam that was published after the cutoff date. The results again indicated that o1-preview performed strongly (97.8th percentile), which suggests that contamination was not a factor. We also show that there is some variability in the output of o1-preview, which means that sometimes there is 'luck' (the answer is correct) or 'bad luck' (the output has diverged into something that is incorrect). We demonstrate that the self-consistency approach, where repeated prompts are given and the most common answer is selected, is a useful strategy for identifying the correct answer. It is concluded that while OpenAI's new model series holds great potential, certain risks must be considered.

Text classification involves categorizing a given text, such as determining its sentiment or identifying harmful content. With the advancement of large language models (LLMs), these models have become highly effective at performing text classification tasks. However, they still show vulnerabilities to variations in text formatting. Recent research demonstrates that modifying input formats, such as vertically aligning words for encoder-based models, can substantially lower accuracy in text classification tasks. While easily understood by humans, these inputs can significantly mislead models, posing a potential risk of bypassing detection in real-world scenarios involving harmful or sensitive information. With the expanding application of LLMs, a crucial question arises: Do decoder-based LLMs exhibit similar vulnerabilities to vertically formatted text input? In this paper, we investigate the impact of vertical text input on the performance of various LLMs across multiple text classification datasets and analyze the underlying causes. Our findings are as follows: (i) Vertical text input significantly degrades the accuracy of LLMs in text classification tasks. (ii) Chain of Thought (CoT) reasoning does not help LLMs recognize vertical input or mitigate its vulnerability, but few-shot learning with careful analysis does. (iii) We explore the underlying cause of the vulnerability by analyzing the inherent issues in tokenization and attention matrices.

Fine-tuning pre-trained large language models (LLMs) on a diverse array of tasks has become a common approach for building models that can solve various natural language processing (NLP) tasks. However, where and to what extent these models retain task-specific knowledge remains largely unexplored. This study investigates the task-specific information encoded in pre-trained LLMs and the effects of instruction tuning on their representations across a diverse set of over 60 NLP tasks. We use a set of matrix analysis tools to examine the differences between the way pre-trained and instruction-tuned LLMs store task-specific information. Our findings reveal that while some tasks are already encoded within the pre-trained LLMs, others greatly benefit from instruction tuning. Additionally, we pinpointed the layers in which the model transitions from high-level general representations to more task-oriented representations. This finding extends our understanding of the governing mechanisms of LLMs and facilitates future research in the fields of parameter-efficient transfer learning and multi-task learning.