This paper details a technical plan for building a clinical case database for Traditional Chinese Medicine (TCM) using web scraping. Leveraging multiple platforms, including 360doc, we gathered over 5,000 TCM clinical cases, performed data cleaning, and structured the dataset with crucial fields such as patient details, pathogenesis, syndromes, and annotations. Using the $Baidu\_ERNIE\_Speed\_128K$ API, we removed redundant information and generated the final answers through the $DeepSeekv2$ API, outputting results in standard JSON format. We optimized data recall with RAG and rerank techniques during retrieval and developed a hybrid matching scheme. By combining two-stage retrieval method with keyword matching via Jieba, we significantly enhanced the accuracy of model outputs.

Large Language Models (LLMs) have strong instruction-following capability to interpret and execute tasks as directed by human commands. Multimodal Large Language Models (MLLMs) have inferior instruction-following ability compared to LLMs. However, there is a significant gap in the instruction-following capabilities between the MLLMs and LLMs. In this study, we conduct a pilot experiment, which demonstrates that spatially down-sampling visual tokens significantly enhances the instruction-following capability of MLLMs. This is attributed to the substantial redundancy in visual modality. However, this intuitive method severely impairs the MLLM's multimodal understanding capability. In this paper, we propose Visual-Modality Token Compression (VMTC) and Cross-Modality Attention Inhibition (CMAI) strategies to alleviate this gap between MLLMs and LLMs by inhibiting the influence of irrelevant visual tokens during content generation, increasing the instruction-following ability of the MLLMs while retaining their multimodal understanding capacity. In VMTC module, the primary tokens are retained and the redundant tokens are condensed by token clustering and merging. In CMAI process, we aggregate text-to-image attentions by text-to-text attentions to obtain a text-to-image focus score. Attention inhibition is performed on the text-image token pairs with low scores. Our comprehensive experiments over instruction-following capabilities and VQA-V2, GQA, TextVQA, MME and MMBench five benchmarks, demonstrate that proposed strategy significantly enhances the instruction following capability of MLLMs while preserving the ability to understand and process multimodal inputs.

Large Language Models (LLMs) have shown to be a great success in a wide range of applications ranging from regular NLP-based use cases to AI agents. LLMs have been trained on a vast corpus of texts from various sources; despite the best efforts during the data pre-processing stage while training the LLMs, they may pick some undesirable information such as personally identifiable information (PII). Consequently, in recent times research in the area of Machine Unlearning (MUL) has become active, the main idea is to force LLMs to forget (unlearn) certain information (e.g., PII) without suffering from performance loss on regular tasks. In this work, we examine the robustness of the existing MUL techniques for their ability to enable leakage-proof forgetting in LLMs. In particular, we examine the effect of data transformation on forgetting, i.e., is an unlearned LLM able to recall forgotten information if there is a change in the format of the input? Our findings on the TOFU dataset highlight the necessity of using diverse data formats to quantify unlearning in LLMs more reliably.

Calibration is crucial as large language models (LLMs) are increasingly deployed to convert natural language queries into SQL for commercial databases. In this work, we investigate calibration techniques for assigning confidence to generated SQL queries. We show that a straightforward baseline -- deriving confidence from the model's full-sequence probability -- outperforms recent methods that rely on follow-up prompts for self-checking and confidence verbalization. Our comprehensive evaluation, conducted across two widely-used Text-to-SQL benchmarks and multiple LLM architectures, provides valuable insights into the effectiveness of various calibration strategies.

Recently, inspired by the concept of sparsity, Mixture-of-Experts (MoE) models have gained increasing popularity for scaling model size while keeping the number of activated parameters constant. In this study, we thoroughly investigate the sparsity of the dense LLaMA model by constructing MoE for both the attention (i.e., Attention MoE) and MLP (i.e., MLP MoE) modules in the transformer blocks. Specifically, we investigate different expert construction methods and granularities under the same activation conditions to analyze the impact of sparsifying the model. Additionally, to comprehensively evaluate the model's capabilities across various domains (e.g., conversation, code, math) after sparsification, we apply sparsity to the instructed large language models (LLMs) and construct instructed MoE models. To counteract the performance degradation resulting from increased sparsity, we design a two-stage post-training strategy to enhance model performance. Experiments on the LLaMA3 model demonstrate the potential effectiveness of this approach for future developments of instructed MoE models. The source codes and models are available at: \url{https://github.com/OpenSparseLLMs/LLaMA-MoE-v2}.

Emerging Large Language Model (LLM) applications require long input prompts to perform complex downstream tasks like document analysis and code generation. For these long context length applications, the length of the input prompt poses a significant challenge in terms of inference efficiency since the inference costs increase linearly with sequence length. However, for many of these applications, much of the context in the prompt is fixed across different user inputs, thereby providing the opportunity to perform offline optimizations to process user inputs quickly, as they are received. In this work, we propose Squeezed Attention as a mechanism to accelerate LLM applications where a large portion of the input prompt is fixed. We first leverage K-means clustering offline to group the keys for the fixed context based on semantic similarity and represent each cluster with a single centroid value. During inference, we compare query tokens from the user input with the centroids to predict which of the keys from the fixed context are semantically relevant and need to be loaded during inference. We then compute exact attention using only these important keys from the fixed context, thereby reducing bandwidth and computational costs. We also extend our method to use a hierarchical centroid lookup to identify important keys, which can reduce the complexity of attention from linear to logarithmic with respect to the context length. We implement optimized Triton kernels for centroid comparison and sparse FlashAttention with important keys, achieving more than 4x speedups during both the prefill and generation phases for long-context inference. Furthermore, we have extensively evaluated our method on various long-context benchmarks including LongBench, where it achieves a 3x reduction in KV cache budget without accuracy loss and up to an 8x reduction with <0.5 point accuracy gap for various models.

Most current molecular language models transfer the masked language model or image-text generation model from natural language processing to molecular field. However, molecules are not solely characterized by atom/bond symbols; they encapsulate important physical/chemical properties. Moreover, normal language models bring grammar rules that are irrelevant for understanding molecules. In this study, we propose a novel physicochemical knowledge-guided molecular meta language framework MolMetaLM. We design a molecule-specialized meta language paradigm, formatted as multiple (subject, predicate, object) knowledge triples sharing the same S (i.e., molecule) to enhance learning the semantic relationships between physicochemical knowledge and molecules. By introducing different molecular knowledge and noises, the meta language paradigm generates tens of thousands of pretraining tasks. By recovering the token/sequence/order-level noises, MolMetaLM exhibits proficiency in large-scale benchmark evaluations involving property prediction, molecule generation, conformation inference, and molecular optimization. Through MolMetaLM, we offer a new insight for designing language models.

Visual generation models have achieved remarkable progress in computer graphics applications but still face significant challenges in real-world deployment. Current assessment approaches for visual generation tasks typically follow an isolated three-phase framework: test input collection, model output generation, and user assessment. These fashions suffer from fixed coverage, evolving difficulty, and data leakage risks, limiting their effectiveness in comprehensively evaluating increasingly complex generation models. To address these limitations, we propose DyEval, an LLM-powered dynamic interactive visual assessment framework that facilitates collaborative evaluation between humans and generative models for text-to-image systems. DyEval features an intuitive visual interface that enables users to interactively explore and analyze model behaviors, while adaptively generating hierarchical, fine-grained, and diverse textual inputs to continuously probe the capability boundaries of the models based on their feedback. Additionally, to provide interpretable analysis for users to further improve tested models, we develop a contextual reflection module that mines failure triggers of test inputs and reflects model potential failure patterns supporting in-depth analysis using the logical reasoning ability of LLM. Qualitative and quantitative experiments demonstrate that DyEval can effectively help users identify max up to 2.56 times generation failures than conventional methods, and uncover complex and rare failure patterns, such as issues with pronoun generation and specific cultural context generation. Our framework provides valuable insights for improving generative models and has broad implications for advancing the reliability and capabilities of visual generation systems across various domains.

Large Language Models (LLMs) offer promising solutions for text summarization. However, some domains require specific information to be available in the summaries. Generating these domain-adapted summaries is still an open challenge. Similarly, hallucinations in generated content is a major drawback of current approaches, preventing their deployment. This study proposes a novel approach that leverages ontologies to create domain-adapted summaries both structured and unstructured. We employ an ontology-guided constrained decoding process to reduce hallucinations while improving relevance. When applied to the medical domain, our method shows potential in summarizing Electronic Health Records (EHRs) across different specialties, allowing doctors to focus on the most relevant information to their domain. Evaluation on the MIMIC-III dataset demonstrates improvements in generating domain-adapted summaries of clinical notes and hallucination reduction.

These models, due to their size--often reaching hundreds of gigabytes--and computational requirements, encounter delays due to what is known as the coldstart This review report discusses the cold start latency in problem [22]. This latency arises when serverless serverless inference and existing solutions. It particularly functions, previously idle, initiate, leading to delays reviews the ServerlessLLM method, a system from the loading of extensive LLM checkpoints designed to address the cold-start problem in serverless and GPU resource activation. Such cold starts can inference for large language models (LLMs). Traditional significantly hinder performance in applications requiring serverless approaches struggle with high latency real-time interaction, making solutions to this due to the size of LLM checkpoints and the problem imperative for scalable, serverless LLM deployment.