Quantization has established itself as the primary approach for decreasing the computational and storage expenses associated with Large Language Models (LLMs) inference. The majority of current research emphasizes quantizing weights and activations to enable low-bit general-matrix-multiply (GEMM) operations, with the remaining non-linear operations executed at higher precision. In our study, we discovered that following the application of these techniques, the primary bottleneck in LLMs inference lies in the softmax layer. The softmax operation comprises three phases: exponent calculation, accumulation, and normalization, Our work focuses on optimizing the first two phases. We propose an analytical approach to determine the optimal clipping value for the input to the softmax function, enabling sub-4-bit quantization for LLMs inference. This method accelerates the calculations of both $e^x$ and $\sum(e^x)$ with minimal to no accuracy degradation. For example, in LLaMA1-30B, we achieve baseline performance with 2-bit quantization on the well-known "Physical Interaction: Question Answering" (PIQA) dataset evaluation. This ultra-low bit quantization allows, for the first time, an acceleration of approximately 4x in the accumulation phase. The combination of accelerating both $e^x$ and $\sum(e^x)$ results in a 36.9% acceleration in the softmax operation.

This study is the first to explore whether multi-modal large language models (LLMs) can align their behaviors with visual personas, addressing a significant gap in the literature that predominantly focuses on text-based personas. We developed a novel dataset of 5K fictional avatar images for assignment as visual personas to LLMs, and analyzed their negotiation behaviors based on the visual traits depicted in these images, with a particular focus on aggressiveness. The results indicate that LLMs assess the aggressiveness of images in a manner similar to humans and output more aggressive negotiation behaviors when prompted with an aggressive visual persona. Interestingly, the LLM exhibited more aggressive negotiation behaviors when the opponent's image appeared less aggressive than their own, and less aggressive behaviors when the opponents image appeared more aggressive.

We show that autoregressive decoding of a transformer-based language model can realize universal computation, without external intervention or modification of the model's weights. Establishing this result requires understanding how a language model can process arbitrarily long inputs using a bounded context. For this purpose, we consider a generalization of autoregressive decoding where, given a long input, emitted tokens are appended to the end of the sequence as the context window advances. We first show that the resulting system corresponds to a classical model of computation, a Lag system, that has long been known to be computationally universal. By leveraging a new proof, we show that a universal Turing machine can be simulated by a Lag system with 2027 production rules. We then investigate whether an existing large language model can simulate the behaviour of such a universal Lag system. We give an affirmative answer by showing that a single system-prompt can be developed for gemini-1.5-pro-001 that drives the model, under deterministic (greedy) decoding, to correctly apply each of the 2027 production rules. We conclude that, by the Church-Turing thesis, prompted gemini-1.5-pro-001 with extended autoregressive (greedy) decoding is a general purpose computer.

In recent years, zero-shot and few-shot learning in visual grounding have garnered considerable attention, largely due to the success of large-scale vision-language pre-training on expansive datasets such as LAION-5B and DataComp-1B. However, the continuous expansion of these datasets presents significant challenges, particularly with respect to data availability and computational overhead, thus creating a bottleneck in the advancement of low-shot learning capabilities. In this paper, we propose IMAGE, Interpretative MAsking with Gaussian radiation modEling, aimed at enhancing vocabulary grounding in low-shot learning scenarios without necessitating an increase in dataset size. Drawing inspiration from cognitive science and the recent success of masked autoencoders (MAE), our method leverages adaptive masking on salient regions of the feature maps generated by the vision backbone. This enables the model to learn robust, generalized representations through the reconstruction of occluded information, thereby facilitating effective attention to both local and global features. We evaluate the efficacy of our approach on benchmark datasets, including COCO and ODinW, demonstrating its superior performance in zero-shot and few-shot tasks. Experimental results consistently show that IMAGE outperforms baseline models, achieving enhanced generalization and improved performance in low-shot scenarios. These findings highlight the potential of adaptive feature manipulation through attention mechanisms and Gaussian modeling as a promising alternative to approaches that rely on the continual scaling of dataset sizes for the advancement of zero-shot and few-shot learning. Our code is publicly available at https://github.com/git-lenny/IMAGE.

Large language models (LLMs) fine-tuned with alignment techniques, such as reinforcement learning from human feedback, have been instrumental in developing some of the most capable AI systems to date. Despite their success, existing methods typically rely on simple binary labels, such as those indicating preferred outputs in pairwise preferences, which fail to capture the subtle differences in relative quality between pairs. To address this limitation, we introduce an approach called Margin Matching Preference Optimization (MMPO), which incorporates relative quality margins into optimization, leading to improved LLM policies and reward models. Specifically, given quality margins in pairwise preferences, we design soft target probabilities based on the Bradley-Terry model, which are then used to train models with the standard cross-entropy objective. Experiments with both human and AI feedback data demonstrate that MMPO consistently outperforms baseline methods, often by a substantial margin, on popular benchmarks including MT-bench and RewardBench. Notably, the 7B model trained with MMPO achieves state-of-the-art performance on RewardBench as of June 2024, outperforming other models of the same scale. Our analysis also shows that MMPO is more robust to overfitting, leading to better-calibrated models.

Large language models exhibit a remarkable capacity for in-context learning, where they learn to solve tasks given a few examples. Recent work has shown that transformers can be trained to perform simple regression tasks in-context. This work explores the possibility of training an in-context learner for classification tasks involving spurious features. We find that the conventional approach of training in-context learners is susceptible to spurious features. Moreover, when the meta-training dataset includes instances of only one task, the conventional approach leads to task memorization and fails to produce a model that leverages context for predictions. Based on these observations, we propose a novel technique to train such a learner for a given classification task. Remarkably, this in-context learner matches and sometimes outperforms strong methods like ERM and GroupDRO. However, unlike these algorithms, it does not generalize well to other tasks. We show that it is possible to obtain an in-context learner that generalizes to unseen tasks by training on a diverse dataset of synthetic in-context learning instances.

Recent advancements in large language models (LLMs) have demonstrated impressive performance in generating molecular structures as drug candidates, which offers significant potential to accelerate drug discovery. However, the current LLMs overlook a critical requirement for drug discovery: proposing a diverse set of molecules. This diversity is essential for improving the chances of finding a viable drug, as it provides alternative molecules that may succeed where others fail in wet-lab or clinical validations. Despite such a need for diversity, the LLMs often output structurally similar molecules from a given prompt. While decoding schemes like beam search may enhance textual diversity, this often does not align with molecular structural diversity. In response, we propose a new method for fine-tuning molecular generative LLMs to autoregressively generate a set of structurally diverse molecules, where each molecule is generated by conditioning on the previously generated molecules. Our approach consists of two stages: (1) supervised fine-tuning to adapt LLMs to autoregressively generate molecules in a sequence and (2) reinforcement learning to maximize structural diversity within the generated molecules. Our experiments show that (1) our fine-tuning approach enables the LLMs to better discover diverse molecules compared to existing decoding schemes and (2) our fine-tuned model outperforms other representative LLMs in generating diverse molecules, including the ones fine-tuned on chemical domains.

Large language models (LLMs) have increased the demand for personalized and stylish content generation. However, closed-source models like GPT-4 present limitations in optimization opportunities, while the substantial training costs and inflexibility of open-source alternatives, such as Qwen-72B, pose considerable challenges. Conversely, small language models (SLMs) struggle with understanding complex instructions and transferring learned capabilities to new contexts, often exhibiting more pronounced limitations. In this paper, we present a novel collaborative training framework that leverages the strengths of both LLMs and SLMs for style article generation, surpassing the performance of either model alone. We freeze the LLMs to harness their robust instruction-following capabilities and subsequently apply supervised fine-tuning on the SLM using style-specific data. Additionally, we introduce a self-improvement method to enhance style consistency. Our new benchmark, NoteBench, thoroughly evaluates style-aligned generation. Extensive experiments show that our approach achieves state-of-the-art performance, with improvements of 0.78 in ROUGE-L and 0.55 in BLEU-4 scores compared to GPT-4, while maintaining a low hallucination rate regarding factual and faithfulness.

Remaining useful life (RUL) prediction is crucial for maintaining modern industrial systems, where equipment reliability and operational safety are paramount. Traditional methods, based on small-scale deep learning or physical/statistical models, often struggle with complex, multidimensional sensor data and varying operating conditions, limiting their generalization capabilities. To address these challenges, this paper introduces an innovative regression framework utilizing large language models (LLMs) for RUL prediction. By leveraging the modelling power of LLMs pre-trained on corpus data, the proposed model can effectively capture complex temporal dependencies and improve prediction accuracy. Extensive experiments on the Turbofan engine's RUL prediction task show that the proposed model surpasses state-of-the-art (SOTA) methods on the challenging FD002 and FD004 subsets and achieves near-SOTA results on the other subsets. Notably, different from previous researches, our framework uses the same sliding window length and all sensor signals for all subsets, demonstrating strong consistency and generalization. Moreover, transfer learning experiments reveal that with minimal target domain data for fine-tuning, the model outperforms SOTA methods trained on full target domain data. This research highlights the significant potential of LLMs in industrial signal processing and RUL prediction, offering a forward-looking solution for health management in future intelligent industrial systems.

Retrieval-Augmented Generation (RAG) has gained significant popularity in modern Large Language Models (LLMs) due to its effectiveness in introducing new knowledge and reducing hallucinations. However, the deep understanding of RAG remains limited, how does RAG help the reasoning process and can RAG help improve the reasoning capability remains question. While external documents are typically considered as a method to incorporate domain-specific information, they also contain intermediate reasoning results related to the query, this suggests that documents could enhance the reasoning capability of LLMs, which has not been previously explored. In this paper, we investigate this issue in depth and find that while RAG can assist with reasoning, the help is limited. If we conceptualize the reasoning process as a tree with fixed depth, then RAG struggles to assist LLMs in performing deeper reasoning. Additionally, the information in the documents requires preprocessing to filter out noise. We demonstrate that this preprocessing is difficult to achieve simply fine-tuning of the LLM, it often necessitates numerous additional transformer layers to solve the problem. To simplify the problem, we propose DPrompt tuning, which effectively resolves the issue within just limited transformer layers, leading to improved performance. Large Language Models (LLMs) (Brown, 2020) have demonstrated remarkable capabilities across a variety of tasks, including text generation and question answering (Ouyang et al., 2022; Wei et al., 2022), code generation (Gu, 2023), and information retrieval (Dai et al., 2024). However, current LLMs often suffer from serious hallucinations (Huang et al., 2023) due to a lack of factual information. Moreover, the knowledge embedded within LLMs is encoded in their parameters (Yang et al., 2024), meaning that incorporating new knowledge requires further fine-tuning, which is both time-consuming and resource-intensive.