Chain of Thought (CoT) is significant in improving the reasoning abilities of large language models (LLMs). However, the correlation between the effectiveness of CoT and the length of reasoning steps in prompts remains largely unknown. To shed light on this, we have conducted several empirical experiments to explore the relations. Specifically, we design experiments that expand and compress the rationale reasoning steps within CoT demonstrations, while keeping all other factors constant. We have the following key findings. First, the results indicate that lengthening the reasoning steps in prompts, even without adding new information into the prompt, considerably enhances LLMs' reasoning abilities across multiple datasets. Alternatively, shortening the reasoning steps, even while preserving the key information, significantly diminishes the reasoning abilities of models. This finding highlights the importance of the number of steps in CoT prompts and provides practical guidance to make better use of LLMs' potential in complex problem-solving scenarios. Second, we also investigated the relationship between the performance of CoT and the rationales used in demonstrations. Surprisingly, the result shows that even incorrect rationales can yield favorable outcomes if they maintain the requisite length of inference. Third, we observed that the advantages of increasing reasoning steps are task-dependent: simpler tasks require fewer steps, whereas complex tasks gain significantly from longer inference sequences.
相關內容
Functional Magnetic Resonance Imaging (fMRI) data is a widely used kind of four-dimensional biomedical data, which requires effective compression. However, fMRI compressing poses unique challenges due to its intricate temporal dynamics, low signal-to-noise ratio, and complicated underlying redundancies. This paper reports a novel compression paradigm specifically tailored for fMRI data based on Implicit Neural Representation (INR). The proposed approach focuses on removing the various redundancies among the time series by employing several methods, including (i) conducting spatial correlation modeling for intra-region dynamics, (ii) decomposing reusable neuronal activation patterns, and (iii) using proper initialization together with nonlinear fusion to describe the inter-region similarity. This scheme appropriately incorporates the unique features of fMRI data, and experimental results on publicly available datasets demonstrate the effectiveness of the proposed method, surpassing state-of-the-art algorithms in both conventional image quality evaluation metrics and fMRI downstream tasks. This work in this paper paves the way for sharing massive fMRI data at low bandwidth and high fidelity.
Pointing out the Shortcomings of Relation Extraction Models with Semantically Motivated Adversarials
In recent years, large language models have achieved state-of-the-art performance across various NLP tasks. However, investigations have shown that these models tend to rely on shortcut features, leading to inaccurate predictions and causing the models to be unreliable at generalization to out-of-distribution (OOD) samples. For instance, in the context of relation extraction (RE), we would expect a model to identify the same relation independently of the entities involved in it. For example, consider the sentence "Leonardo da Vinci painted the Mona Lisa" expressing the created(Leonardo_da_Vinci, Mona_Lisa) relation. If we substiute "Leonardo da Vinci" with "Barack Obama", then the sentence still expresses the created relation. A robust model is supposed to detect the same relation in both cases. In this work, we describe several semantically-motivated strategies to generate adversarial examples by replacing entity mentions and investigate how state-of-the-art RE models perform under pressure. Our analyses show that the performance of these models significantly deteriorates on the modified datasets (avg. of -48.5% in F1), which indicates that these models rely to a great extent on shortcuts, such as surface forms (or patterns therein) of entities, without making full use of the information present in the sentences.
With the advent of large language models (LLMs) like GPT-3, a natural question is the extent to which these models can be utilized for source code optimization. This paper presents methodologically stringent case studies applied to well-known open source python libraries pillow and numpy. We find that contemporary LLM ChatGPT-4 (state September and October 2023) is surprisingly adept at optimizing energy and compute efficiency. However, this is only the case in interactive use, with a human expert in the loop. Aware of experimenter bias, we document our qualitative approach in detail, and provide transcript and source code. We start by providing a detailed description of our approach in conversing with the LLM to optimize the _getextrema function in the pillow library, and a quantitative evaluation of the performance improvement. To demonstrate qualitative replicability, we report further attempts on another locus in the pillow library, and one code locus in the numpy library, to demonstrate generalization within and beyond a library. In all attempts, the performance improvement is significant (factor up to 38). We have also not omitted reporting of failed attempts (there were none). We conclude that LLMs are a promising tool for code optimization in open source libraries, but that the human expert in the loop is essential for success. Nonetheless, we were surprised by how few iterations were required to achieve substantial performance improvements that were not obvious to the expert in the loop. We would like bring attention to the qualitative nature of this study, more robust quantitative studies would need to introduce a layer of selecting experts in a representative sample -- we invite the community to collaborate.
Recently, the advent of Large Visual-Language Models (LVLMs) has received increasing attention across various domains, particularly in the field of visual document understanding (VDU). Different from conventional vision-language tasks, VDU is specifically concerned with text-rich scenarios containing abundant document elements. Nevertheless, the importance of fine-grained features remains largely unexplored within the community of LVLMs, leading to suboptimal performance in text-rich scenarios. In this paper, we abbreviate it as the fine-grained feature collapse issue. With the aim of filling this gap, we propose a contrastive learning framework, termed Document Object COntrastive learning (DoCo), specifically tailored for the downstream tasks of VDU. DoCo leverages an auxiliary multimodal encoder to obtain the features of document objects and align them to the visual features generated by the vision encoder of LVLM, which enhances visual representation in text-rich scenarios. It can represent that the contrastive learning between the visual holistic representations and the multimodal fine-grained features of document objects can assist the vision encoder in acquiring more effective visual cues, thereby enhancing the comprehension of text-rich documents in LVLMs. We also demonstrate that the proposed DoCo serves as a plug-and-play pre-training method, which can be employed in the pre-training of various LVLMs without inducing any increase in computational complexity during the inference process. Extensive experimental results on multiple benchmarks of VDU reveal that LVLMs equipped with our proposed DoCo can achieve superior performance and mitigate the gap between VDU and generic vision-language tasks.
We introduce two algorithms for computing tight guarantees on the probabilistic robustness of Bayesian Neural Networks (BNNs). Computing robustness guarantees for BNNs is a significantly more challenging task than verifying the robustness of standard Neural Networks (NNs) because it requires searching the parameters' space for safe weights. Moreover, tight and complete approaches for the verification of standard NNs, such as those based on Mixed-Integer Linear Programming (MILP), cannot be directly used for the verification of BNNs because of the polynomial terms resulting from the consecutive multiplication of variables encoding the weights. Our algorithms efficiently and effectively search the parameters' space for safe weights by using iterative expansion and the network's gradient and can be used with any verification algorithm of choice for BNNs. In addition to proving that our algorithms compute tighter bounds than the SoA, we also evaluate our algorithms against the SoA on standard benchmarks, such as MNIST and CIFAR10, showing that our algorithms compute bounds up to 40% tighter than the SoA.
Given the emergent reasoning abilities of large language models, information retrieval is becoming more complex. Rather than just retrieve a document, modern information retrieval systems advertise that they can synthesize an answer based on potentially many different documents, conflicting data sources, and using reasoning. But, different kinds of questions have different answers, and different answers have different complexities. In this paper, we introduce a novel framework for analyzing the complexity of a question answer based on the natural deduction calculus as presented in Prawitz (1965). Our framework is novel both in that no one to our knowledge has used this logic as a basis for complexity classes, and also in that no other existing complexity classes to these have been delineated using any analogous methods either. We identify three decidable fragments in particular called the forward, query and planning fragments, and we compare this to what would be needed to do proofs for the complete first-order calculus, for which theorem-proving is long known to be undecidable.
Instruction Tuning (IT), the process of training large language models (LLMs) using instruction-response pairs, has emerged as the predominant method for transforming base pre-trained LLMs into open-domain conversational agents. While IT has achieved notable success and widespread adoption, its limitations and shortcomings remain underexplored. In this paper, through rigorous experiments and an in-depth analysis of the changes LLMs undergo through IT, we reveal various limitations of IT. In particular, we show that (1) IT fails to enhance knowledge or skills in LLMs. LoRA fine-tuning is limited to learning response initiation and style tokens, and full-parameter fine-tuning leads to knowledge degradation. (2) Copying response patterns from IT datasets derived from knowledgeable sources leads to a decline in response quality. (3) Full-parameter fine-tuning increases hallucination by inaccurately borrowing tokens from conceptually similar instances in the IT dataset for generating responses. (4) Popular methods to improve IT do not lead to performance improvements over a simple LoRA fine-tuned model. Our findings reveal that responses generated solely from pre-trained knowledge consistently outperform responses by models that learn any form of new knowledge from IT on open-source datasets. We hope the insights and challenges revealed inspire future work.
Model editing aims to precisely modify the behaviours of large language models (LLMs) on specific knowledge while keeping irrelevant knowledge unchanged. It has been proven effective in resolving hallucination and out-of-date issues in LLMs. As a result, it can boost the application of LLMs in many critical domains (e.g., medical domain), where the hallucination is not tolerable. In this paper, we propose two model editing studies and validate them in the medical domain: (1) directly editing the factual medical knowledge and (2) editing the explanations to facts. Meanwhile, we observed that current model editing methods struggle with the specialization and complexity of medical knowledge. Therefore, we propose MedLaSA, a novel Layer-wise Scalable Adapter strategy for medical model editing. It employs causal tracing to identify the precise location of knowledge in neurons and then introduces scalable adapters into the dense layers of LLMs. These adapters are assigned scaling values based on the corresponding specific knowledge. To evaluate the editing impact, we build two benchmark datasets and introduce a series of challenging and comprehensive metrics. Extensive experiments on medical LLMs demonstrate the editing efficiency of MedLaSA, without affecting irrelevant knowledge that is not edited.
Large Language Models (LLMs) represent an advanced evolution of earlier, simpler language models. They boast enhanced abilities to handle complex language patterns and generate coherent text, images, audios, and videos. Furthermore, they can be fine-tuned for specific tasks. This versatility has led to the proliferation and extensive use of numerous commercialized large models. However, the rapid expansion of LLMs has raised security and ethical concerns within the academic community. This emphasizes the need for ongoing research into security evaluation during their development and deployment. Over the past few years, a substantial body of research has been dedicated to the security evaluation of large-scale models. This article an in-depth review of the most recent advancements in this field, providing a comprehensive analysis of commonly used evaluation metrics, advanced evaluation frameworks, and the routine evaluation processes for LLMs. Furthermore, we also discuss the future directions for advancing the security evaluation of LLMs.
Named entity recognition (NER) is the task to identify text spans that mention named entities, and to classify them into predefined categories such as person, location, organization etc. NER serves as the basis for a variety of natural language applications such as question answering, text summarization, and machine translation. Although early NER systems are successful in producing decent recognition accuracy, they often require much human effort in carefully designing rules or features. In recent years, deep learning, empowered by continuous real-valued vector representations and semantic composition through nonlinear processing, has been employed in NER systems, yielding stat-of-the-art performance. In this paper, we provide a comprehensive review on existing deep learning techniques for NER. We first introduce NER resources, including tagged NER corpora and off-the-shelf NER tools. Then, we systematically categorize existing works based on a taxonomy along three axes: distributed representations for input, context encoder, and tag decoder. Next, we survey the most representative methods for recent applied techniques of deep learning in new NER problem settings and applications. Finally, we present readers with the challenges faced by NER systems and outline future directions in this area.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191