Haiyun Jiang

Longyue Wang, Zefeng Du, Donghuai Liu et al.

Modeling discourse -- the linguistic phenomena that go beyond individual sentences, is a fundamental yet challenging aspect of natural language processing (NLP). However, existing evaluation benchmarks primarily focus on the evaluation of inter-sentence properties and overlook critical discourse phenomena that cross sentences. To bridge the gap, we propose Disco-Bench, a benchmark that can evaluate intra-sentence discourse properties across a diverse set of NLP tasks, covering understanding, translation, and generation. Disco-Bench consists of 9 document-level testsets in the literature domain, which contain rich discourse phenomena (e.g. cohesion and coherence) in Chinese and/or English. For linguistic analysis, we also design a diagnostic test suite that can examine whether the target models learn discourse knowledge. We totally evaluate 20 general-, in-domain and commercial models based on Transformer, advanced pretraining architectures and large language models (LLMs). Our results show (1) the challenge and necessity of our evaluation benchmark; (2) fine-grained pretraining based on literary document-level training data consistently improves the modeling of discourse information. We will release the datasets, pretrained models, and leaderboard, which we hope can significantly facilitate research in this field: https://github.com/longyuewangdcu/Disco-Bench.

Yi Chen, Rui Wang, Haiyun Jiang et al.

Evaluating the quality of generated text is a challenging task in NLP, due to the inherent complexity and diversity of text. Recently, large language models (LLMs) have garnered significant attention due to their impressive performance in various tasks. Therefore, we present this paper to investigate the effectiveness of LLMs, especially ChatGPT, and explore ways to optimize their use in assessing text quality. We compared three kinds of reference-free evaluation methods. The experimental results prove that ChatGPT is capable of evaluating text quality effectively from various perspectives without reference and demonstrates superior performance than most existing automatic metrics. In particular, the Explicit Score, which utilizes ChatGPT to generate a numeric score measuring text quality, is the most effective and reliable method among the three exploited approaches. However, directly comparing the quality of two texts may lead to suboptimal results. We believe this paper will provide valuable insights for evaluating text quality with LLMs and have released the used data.

Mingyang Song, Haiyun Jiang, Shuming Shi et al.

The emergence of ChatGPT has recently garnered significant attention from the computational linguistics community. To demonstrate its capabilities as a keyphrase generator, we conduct a preliminary evaluation of ChatGPT for the keyphrase generation task. We evaluate its performance in various aspects, including keyphrase generation prompts, keyphrase generation diversity, and long document understanding. Our evaluation is based on six benchmark datasets, and we adopt the prompt suggested by OpenAI while extending it to six candidate prompts. We find that ChatGPT performs exceptionally well on all six candidate prompts, with minor performance differences observed across the datasets. Based on our findings, we conclude that ChatGPT has great potential for keyphrase generation. Moreover, we discover that ChatGPT still faces challenges when it comes to generating absent keyphrases. Meanwhile, in the final section, we also present some limitations and future expansions of this report.

Yongrui Chen, Haiyun Jiang, Xinting Huang et al.

The improvement of LLMs' instruction-following capabilities relies heavily on the availability of high-quality instruction-response pairs. Unfortunately, the current methods used to collect the pairs suffer from either unaffordable labor costs or severe hallucinations in the self-generation of LLM. To tackle these challenges, this paper proposes a scalable solution. It involves training LLMs to generate instruction-response pairs based on human-written documents, rather than relying solely on self-generation without context. Our proposed method not only exploits the advantages of human-written documents in reducing hallucinations but also utilizes an LLM to wrap the expression of documents, which enables us to bridge the gap between various document styles and the standard AI response. Experiments demonstrate that our method outperforms existing typical methods on multiple benchmarks. In particular, compared to the best-performing baseline, the LLM trained using our generated dataset exhibits a 10\% relative improvement in performance on AlpacaEval, despite utilizing only 1/5 of its training data. Furthermore, a comprehensive manual evaluation validates the quality of the data we generated. Our trained wrapper is publicly available at https://github.com/Bahuia/Dog-Instruct.

Jingping Liu, Yuqiu Song, Kui Xue et al.

Prompt tuning is an emerging way of adapting pre-trained language models to downstream tasks. However, the existing studies are mainly to add prompts to the input sequence. This way would not work as expected due to the intermediate multi-head self-attention and feed-forward network computation, making model optimization not very smooth. Hence, we propose a novel tuning way called layer tuning, aiming to add learnable parameters in Transformer layers. Specifically, we focus on layer tuning for feed-forward network in the Transformer, namely FL-tuning. It introduces additional units into the hidden layer of each feed-forward network. We conduct extensive experiments on the public CLUE benchmark. The results show that: 1) Our FL-tuning outperforms prompt tuning methods under both full-data and few-shot settings in almost all cases. In particular, it improves accuracy by 17.93% (full-data setting) on WSC 1.0 and F1 by 16.142% (few-shot setting) on CLUENER over P-tuning v2. 2) Our FL-tuning is more stable and converges about 1.17 times faster than P-tuning v2. 3) With only about 3% of Transformer's parameters to be trained, FL-tuning is comparable with fine-tuning on most datasets, and significantly outperforms fine-tuning (e.g., accuracy improved by 12.9% on WSC 1.1) on several datasets. The source codes are available at https://github.com/genggui001/FL-Tuning.

Nuo Chen, Yan Wang, Haiyun Jiang et al.

In recent years, Dialogue-style Large Language Models (LLMs) such as ChatGPT and GPT4 have demonstrated immense potential in constructing open-domain dialogue agents. However, aligning these agents with specific characters or individuals remains a considerable challenge due to the complexities of character representation and the lack of comprehensive annotations. In this paper, we introduce the Harry Potter Dialogue (HPD) dataset, designed to advance the study of dialogue agents and character alignment. The dataset encompasses all dialogue sessions (in both English and Chinese) from the Harry Potter series and is annotated with vital background information, including dialogue scenes, speakers, character relationships, and attributes. These extensive annotations may empower LLMs to unlock character-driven dialogue capabilities. Furthermore, it can serve as a universal benchmark for evaluating how well can a LLM aligning with a specific character. We benchmark LLMs on HPD using both fine-tuning and in-context learning settings. Evaluation results reveal that although there is substantial room for improvement in generating high-quality, character-aligned responses, the proposed dataset is valuable in guiding models toward responses that better align with the character of Harry Potter.

Hongzhan Lin, Pengyao Yi, Jing Ma et al.

The spread of rumors along with breaking events seriously hinders the truth in the era of social media. Previous studies reveal that due to the lack of annotated resources, rumors presented in minority languages are hard to be detected. Furthermore, the unforeseen breaking events not involved in yesterday's news exacerbate the scarcity of data resources. In this work, we propose a novel zero-shot framework based on prompt learning to detect rumors falling in different domains or presented in different languages. More specifically, we firstly represent rumor circulated on social media as diverse propagation threads, then design a hierarchical prompt encoding mechanism to learn language-agnostic contextual representations for both prompts and rumor data. To further enhance domain adaptation, we model the domain-invariant structural features from the propagation threads, to incorporate structural position representations of influential community response. In addition, a new virtual response augmentation method is used to improve model training. Extensive experiments conducted on three real-world datasets demonstrate that our proposed model achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.

Shuming Shi, Enbo Zhao, Duyu Tang et al.

In this technical report, we introduce Effidit (Efficient and Intelligent Editing), a digital writing assistant that facilitates users to write higher-quality text more efficiently by using artificial intelligence (AI) technologies. Previous writing assistants typically provide the function of error checking (to detect and correct spelling and grammatical errors) and limited text-rewriting functionality. With the emergence of large-scale neural language models, some systems support automatically completing a sentence or a paragraph. In Effidit, we significantly expand the capacities of a writing assistant by providing functions in five categories: text completion, error checking, text polishing, keywords to sentences (K2S), and cloud input methods (cloud IME). In the text completion category, Effidit supports generation-based sentence completion, retrieval-based sentence completion, and phrase completion. In contrast, many other writing assistants so far only provide one or two of the three functions. For text polishing, we have three functions: (context-aware) phrase polishing, sentence paraphrasing, and sentence expansion, whereas many other writing assistants often support one or two functions in this category. The main contents of this report include major modules of Effidit, methods for implementing these modules, and evaluation results of some key methods.

Yifan Wang, Qingyan Guo, Xinzhe Ni et al.

In-context learning (ICL) ability has emerged with the increasing scale of large language models (LLMs), enabling them to learn input-label mappings from demonstrations and perform well on downstream tasks. However, under the standard ICL setting, LLMs may sometimes neglect query-related information in demonstrations, leading to incorrect predictions. To address this limitation, we propose a new paradigm called Hint-enhanced In-Context Learning (HICL) to explore the power of ICL in open-domain question answering, an important form in knowledge-intensive tasks. HICL leverages LLMs' reasoning ability to extract query-related knowledge from demonstrations, then concatenates the knowledge to prompt LLMs in a more explicit way. Furthermore, we track the source of this knowledge to identify specific examples, and introduce a Hint-related Example Retriever (HER) to select informative examples for enhanced demonstrations. We evaluate HICL with HER on 3 open-domain QA benchmarks, and observe average performance gains of 2.89 EM score and 2.52 F1 score on gpt-3.5-turbo, 7.62 EM score and 7.27 F1 score on LLaMA-2-Chat-7B compared with standard setting.

Mao Yan Chen, Haiyun Jiang, Yujiu Yang

The short text matching task employs a model to determine whether two short texts have the same semantic meaning or intent. Existing short text matching models usually rely on the content of short texts which are lack information or missing some key clues. Therefore, the short texts need external knowledge to complete their semantic meaning. To address this issue, we propose a new short text matching framework for introducing external knowledge to enhance the short text contextual representation. In detail, we apply a self-attention mechanism to enrich short text representation with external contexts. Experiments on two Chinese datasets and one English dataset demonstrate that our framework outperforms the state-of-the-art short text matching models.

Lingfeng Shen, Haiyun Jiang, Lemao Liu et al.

Sentence embedding is one of the most fundamental tasks in Natural Language Processing and plays an important role in various tasks. The recent breakthrough in sentence embedding is achieved by pre-trained language models (PLMs). Despite its success, an embedded vector (Sen2Vec) representing a point estimate does not naturally express uncertainty in a taskagnostic way. This paper thereby proposes an efficient framework on probabilistic sentence embedding (Sen2Pro) from PLMs, and it represents a sentence as a probability density distribution in an embedding space to reflect both model uncertainty and data uncertainty (i.e., many-to-one nature) in the sentence representation. The proposed framework performs in a plug-and-play way without retraining PLMs anymore, and it is easy to implement and generally applied on top of any PLM. The superiority of Sen2Pro over Sen2Vec has been theoretically verified and practically illustrated on different NLP tasks.

Chaoyang Wang, Zeyu Zhang, Meng Meng et al.

Visual reasoning is crucial for understanding complex multimodal data and advancing Artificial General Intelligence. Existing methods enhance the reasoning capability of Multimodal Large Language Models (MLLMs) through Reinforcement Learning (RL) fine-tuning (e.g., GRPO). However, current RL approaches sample action groups solely from the policy model itself, which limits the upper boundary of the model's reasoning capability and leads to inefficient training. To address these limitations, this paper proposes a novel RL framework called \textbf{Vision-EKIPL}. The core of this framework lies in introducing high-quality actions generated by external auxiliary models during the RL training process to guide the optimization of the policy model. The policy learning with knowledge infusion from external models significantly expands the model's exploration space, effectively improves the reasoning boundary, and substantially accelerates training convergence speed and efficiency. Experimental results demonstrate that our proposed Vision-EKIPL achieved up to a 5\% performance improvement on the Reason-RFT-CoT Benchmark compared to the state-of-the-art (SOTA). It reveals that Vision-EKIPL can overcome the limitations of traditional RL methods, significantly enhance the visual reasoning performance of MLLMs, and provide a new effective paradigm for research in this field.

Chang Gao, Haiyun Jiang, Deng Cai et al.

Most existing prompting methods suffer from the issues of generalizability and consistency, as they often rely on instance-specific solutions that may not be applicable to other instances and lack task-level consistency across the selected few-shot examples. To address these limitations, we propose a comprehensive framework, StrategyLLM, allowing LLMs to perform inductive reasoning, deriving general strategies from specific task instances, and deductive reasoning, applying these general strategies to particular task examples, for constructing generalizable and consistent few-shot prompts. It employs four LLM-based agents: strategy generator, executor, optimizer, and evaluator, working together to generate, evaluate, and select promising strategies for a given task. Experimental results demonstrate that StrategyLLM outperforms the competitive baseline CoT-SC that requires human-annotated solutions on 13 datasets across 4 challenging tasks without human involvement, including math reasoning (34.2\% $\rightarrow$ 38.8\%), commonsense reasoning (70.3\% $\rightarrow$ 72.5\%), algorithmic reasoning (73.7\% $\rightarrow$ 85.0\%), and symbolic reasoning (30.0\% $\rightarrow$ 79.2\%). Further analysis reveals that StrategyLLM is applicable to various LLMs and demonstrates advantages across numerous scenarios.

Lingfeng Shen, Ze Zhang, Haiyun Jiang et al.

Adversarial attack serves as a major challenge for neural network models in NLP, which precludes the model's deployment in safety-critical applications. A recent line of work, detection-based defense, aims to distinguish adversarial sentences from benign ones. However, {the core limitation of previous detection methods is being incapable of giving correct predictions on adversarial sentences unlike defense methods from other paradigms.} To solve this issue, this paper proposes TextShield: (1) we discover a link between text attack and saliency information, and then we propose a saliency-based detector, which can effectively detect whether an input sentence is adversarial or not. (2) We design a saliency-based corrector, which converts the detected adversary sentences to benign ones. By combining the saliency-based detector and corrector, TextShield extends the detection-only paradigm to a detection-correction paradigm, thus filling the gap in the existing detection-based defense. Comprehensive experiments show that (a) TextShield consistently achieves higher or comparable performance than state-of-the-art defense methods across various attacks on different benchmarks. (b) our saliency-based detector outperforms existing detectors for detecting adversarial sentences.

Bolun Sun, Yifan Zhou, Haiyun Jiang

This paper presents a novel application of large language models (LLMs) to enhance user comprehension of privacy policies through an interactive dialogue agent. We demonstrate that LLMs significantly outperform traditional models in tasks like Data Practice Identification, Choice Identification, Policy Summarization, and Privacy Question Answering, setting new benchmarks in privacy policy analysis. Building on these findings, we introduce an innovative LLM-based agent that functions as an expert system for processing website privacy policies, guiding users through complex legal language without requiring them to pose specific questions. A user study with 100 participants showed that users assisted by the agent had higher comprehension levels (mean score of 2.6 out of 3 vs. 1.8 in the control group), reduced cognitive load (task difficulty ratings of 3.2 out of 10 vs. 7.8), increased confidence in managing privacy, and completed tasks in less time (5.5 minutes vs. 15.8 minutes). This work highlights the potential of LLM-based agents to transform user interaction with privacy policies, leading to more informed consent and empowering users in the digital services landscape.

Jincheng Dai, Zhuowei Huang, Haiyun Jiang et al.

Large Language Models (LLMs), despite their remarkable performance across a wide range of tasks, necessitate substantial GPU memory and consume significant computational resources. Beyond the memory taken up by model weights, the memory used by the KV cache rises linearly with sequence length, becoming a primary bottleneck for inference. In this paper, we introduce an innovative method for optimizing the KV cache, which considerably minimizes its memory footprint. Upon thorough investigation, we discover that in most Transformer models, (i) there is a striking similarity between adjacent tokens' query vectors, and (ii) the attention calculation of the current query can rely exclusively on the attention information of a small fraction of preceding queries. Based on these observations, we present CORM, a KV cache eviction policy that dynamically retains essential key-value pairs for inference without the need for model fine-tuning. Our validation shows that CORM reduces the inference memory usage of KV cache by up to 70\% with negligible performance degradation across six tasks in LongBench. Furthermore, we demonstrate that CORM is compatible with GQA for further compression rate.

Qihang Ai, Haiyun Jiang

We study reasoning tasks through a framework that integrates auto-regressive (AR) and non-autoregressive (NAR) language models. AR models, which generate text sequentially, excel at producing coherent outputs but often suffer from slow inference, particularly in reasoning-intensive domains such as mathematics and code, where lengthy chains of thought are required. In contrast, NAR models, such as discrete diffusion models, allow parallel generation and offer substantial speedups, though typically at the cost of reduced output quality. To address these limitations, we introduce a new paradigm in which an NAR model efficiently produces intermediate reasoning traces, which subsequently guide an AR model to deliver precise final answers. Experiments demonstrate that our approach yields significant 26% improvements over strong baselines while substantially reducing inference cost.

Lingfeng Shen, Haiyun Jiang, Lemao Liu et al.

Static word embedding is still useful, particularly for context-unavailable tasks, because in the case of no context available, pre-trained language models often perform worse than static word embeddings. Although dimension is a key factor determining the quality of static word embeddings, automatic dimension selection is rarely discussed. In this paper, we investigate the impact of word frequency on the dimension selection, and empirically find that word frequency is so vital that it needs to be taken into account during dimension selection. Based on such an empirical finding, this paper proposes a dimension selection method that uses a metric (Mixed Product Distance, MPD) to select a proper dimension for word embedding algorithms without training any word embedding. Through applying a post-processing function to oracle matrices, the MPD-based method can de-emphasize the impact of word frequency. Experiments on both context-unavailable and context-available tasks demonstrate the better efficiency-performance trade-off of our MPD-based dimension selection method over baselines.

Lingfeng Shen, Haiyun Jiang, Lemao Liu et al.

Generating proper embedding of sentences through an unsupervised way is beneficial to semantic matching and retrieval problems in real-world scenarios. This paper presents Representation ALchemy (RepAL), an extremely simple post-processing method that enhances sentence representations. The basic idea in RepAL is to de-emphasize redundant information of sentence embedding generated by pre-trained models. Through comprehensive experiments, we show that RepAL is free of training and is a plug-and-play method that can be combined with most existing unsupervised sentence learning models. We also conducted in-depth analysis to understand RepAL.

Lingfeng Shen, Lemao Liu, Haiyun Jiang et al.

In this paper we revisit automatic metrics for paraphrase evaluation and obtain two findings that disobey conventional wisdom: (1) Reference-free metrics achieve better performance than their reference-based counterparts. (2) Most commonly used metrics do not align well with human annotation. Underlying reasons behind the above findings are explored through additional experiments and in-depth analyses. Based on the experiments and analyses, we propose ParaScore, a new evaluation metric for paraphrase generation. It possesses the merits of reference-based and reference-free metrics and explicitly models lexical divergence. Experimental results demonstrate that ParaScore significantly outperforms existing metrics.

Lingfeng Shen, Haiyun Jiang, Lemao Liu et al.

It has been shown that natural language processing (NLP) models are vulnerable to a kind of security threat called the Backdoor Attack, which utilizes a `backdoor trigger' paradigm to mislead the models. The most threatening backdoor attack is the stealthy backdoor, which defines the triggers as text style or syntactic. Although they have achieved an incredible high attack success rate (ASR), we find that the principal factor contributing to their ASR is not the `backdoor trigger' paradigm. Thus the capacity of these stealthy backdoor attacks is overestimated when categorized as backdoor attacks. Therefore, to evaluate the real attack power of backdoor attacks, we propose a new metric called attack successful rate difference (ASRD), which measures the ASR difference between clean state and poison state models. Besides, since the defenses against stealthy backdoor attacks are absent, we propose Trigger Breaker, consisting of two too simple tricks that can defend against stealthy backdoor attacks effectively. Experiments show that our method achieves significantly better performance than state-of-the-art defense methods against stealthy backdoor attacks.