ACL 2024 论文列表

Content moderators play a key role in keeping the conversation on social media healthy. While the high volume of content they need to judge represents a bottleneck to the moderation pipeline, no studies have explored how models could support them to make faster decisions. There is, by now, a vast body of research into detecting hate speech, sometimes explicitly motivated by a desire to help improve content moderation, but published research using real content moderators is scarce. In this work we investigate the effect of explanations on the speed of real-world moderators. Our experiments show that while generic explanations do not affect their speed and are often ignored, structured explanations lower moderators’ decision making time by 7.4%.

Recent advancements in large language models (LLMs) have brought significant changes to various domains, especially through LLM-driven autonomous agents. A representative scenario is in software development, where LLM agents demonstrate efficient collaboration, task division, and assurance of software quality, markedly reducing the need for manual involvement. However, these agents frequently perform a variety of tasks independently, without benefiting from past experiences, which leads to repeated mistakes and inefficient attempts in multi-step task execution. To this end, we introduce Experiential Co-Learning, a novel LLM-agent learning framework in which instructor and assistant agents gather shortcut-oriented experiences from their historical trajectories and use these past experiences for future task execution. The extensive experiments demonstrate that the framework enables agents to tackle unseen software-developing tasks more effectively. We anticipate that our insights will guide LLM agents towards enhanced autonomy and contribute to their evolutionary growth in cooperative learning. The code and data are available at https://github.com/OpenBMB/ChatDev.

We introduce EVLGen, a streamlined framework designed for the pre-training of visually conditioned language generation models with high computational demands, utilizing frozen pre-trained large language models (LLMs). The conventional approach in vision-language pre-training (VLP) typically involves a two-stage optimization process: an initial resource-intensive phase dedicated to general-purpose vision-language representation learning, focused on extracting and consolidating relevant visual features. This is followed by a subsequent phase that emphasizes end-to-end alignment between visual and linguistic modalities. Our novel one-stage, single-loss framework bypasses the computationally demanding first training stage by gradually merging similar visual tokens during training, while avoiding model collapse caused by single-stage training of BLIP-2 type models. The gradual merging process effectively condenses visual information while preserving semantic richness, resulting in rapid convergence without compromising performance. Our experimental findings demonstrate that our approach accelerates the training of vision-language models by a factor of 5 without a noticeable impact on overall performance. Furthermore, we illustrate that our models significantly narrow the performance gap to current vision-language models using only 1/10 of the data. Finally, we showcase how our image-text models can seamlessly adapt to video-conditioned language generation tasks through novel soft attentive temporal token contextualizing modules. Code: https://github.com/yiren-jian/EVLGen

Through the rise of social media platforms, longitudinal language modelling has received much attention over the latest years, especially in downstream tasks such as mental health monitoring of individuals where modelling linguistic content in a temporal fashion is crucial. A key limitation in existing work is how to effectively model temporal sequences within Transformer-based language models. In this work we address this challenge by introducing a novel approach for predicting ‘Moments of Change’ (MoC) in the mood of online users, by simultaneously considering user linguistic and time-aware context. A Hawkes process-inspired transformation layer is applied over the proposed architecture to model the influence of time on users’ posts – capturing both their immediate and historical dynamics. We perform experiments on the two existing datasets for the MoC task and showcase clear performance gains when leveraging the proposed layer. Our ablation study reveals the importance of considering temporal dynamics in detecting subtle and rare mood changes. Our results indicate that considering linguistic and temporal information in a hierarchical manner provide valuable insights into the temporal dynamics of modelling user generated content over time, with applications in mental health monitoring.

Benchmarks have emerged as the central approach for evaluating Large Language Models (LLMs). The research community often relies on a model’s average performance across the test prompts of a benchmark to evaluate the model’s performance. This is consistent with the assumption that the test prompts within a benchmark represent a random sample from some real-world distribution of interest. We note that this is generally not the case; instead, we hold that the distribution of interest varies according to the specific use case. Hence, we analyze the robustness of LLM benchmarks to their underlying distributional assumptions. We find that (1) the correlation in model performance across test prompts is non-random, (2) accounting for correlations across test prompts can change model rankings on major benchmarks, (3) explanatory factors for these correlations include semantic similarity and common LLM failure points.

Generating long-term texts such as novels using artificial intelligence has always been a challenge. A common approach is to use large language models (LLMs) to construct a hierarchical framework that first plans and then writes. Despite the fact that the generated novels reach a sufficient length, they exhibit poor logical coherence and appeal in their plots and deficiencies in character and event depiction, ultimately compromising the overall narrative quality. In this paper, we propose a method named Extracting Excelsior and Expanding. Ex3 initially extract structural information by learning from raw novel data. By combining this structure information with the novel data, an instruction-following dataset is meticulously crafted. This dataset is then utilized to fine-tune the LLM, aiming for excelsior generation performance. In the final stage, a tree-like expansion method is deployed to facilitate the generation of arbitrarily long novels.Evaluation against previous methods showcases Ex3’s ability to produce higher-quality long-form novels.

Current fact verification methods generally follow the two-stage training paradigm: evidence retrieval and claim verification. While existing works focus on developing sophisticated claim verification modules, the fundamental importance of evidence retrieval is largely ignored. Existing approaches usually adopt the heuristic semantic similarity-based retrieval strategy, resulting in the task-irrelevant evidence and undesirable performance. In this paper, we concentrate on evidence retrieval and propose a Retrieval-Augmented Verification framework RAV, consisting of two major modules: the hybrid evidence retrieval and the joint fact verification. Hybrid evidence retrieval module incorporates an efficient retriever for preliminary pruning of candidate evidence, succeeded by a ranker that generates more precise sorting results. Under this end-to-end training paradigm, gradients from the claim verification can be back-propagated to enhance evidence selection. Experimental results on FEVER dataset demonstrate the superiority of RAV.

This paper introduce a novel thought prompting approach called ”Everything of Thoughts” (XoT) for Large Language Models (LLMs) to defy the law of ”Penrose triangle” of existing thought paradigms, to achieve three key perspectives in thought generation simultaneously: performance, efficiency, and flexibility. XoT leverages pretrained reinforcement learning and Monte Carlo Tree Search (MCTS) to incorporate external domain knowledge and planning capability into thoughts, thereby enhancing LLMs’ decision-making capabilities. Through the MCTS-LLM collaborative thought revision framework, XoT autonomously produces high-quality comprehensive cognitive mappings with minimal LLM interactions. Additionally, XoT empowers LLMs to utilize flexible cognitive mappings for solving problems with multiple solutions.We evaluate XoT on several challenging problem-solving tasks, including Game of 24, 8-Puzzle, and Pocket Cube. Our results demonstrate that XoT significantly outperforms existing approaches in various dimensions, showcasing its remarkable proficiency in addressing complex problems across diverse domains. The data and code are available at https://github.com/microsoft/Everything-of-Thoughts-XoT.

Large language models have demonstrated impressive performance on commonsense tasks; however, these tasks are often posed as multiple-choice questions, allowing models to exploit systematic biases. Commonsense is also inherently probabilistic with multiple correct answers. The purpose of “boiling water” could be making tea, cooking but also could be killing germs. Existing tasks do not capture the probabilistic nature of common sense. To this end, we present commonsense frame completion (CFC), a new generative task that evaluates common sense via multiple open-ended generations. We also propose a method of probabilistic evaluation that strongly correlates with human judgments. Humans drastically outperform strong language model baselines on our dataset, indicating this approach is both a challenging and useful evaluation of machine common sense.

The swift detection of multimedia fake news has emerged as a crucial task in combating malicious propaganda and safeguarding the security of the online environment. While existing methods have achieved commendable results in modeling entity-level inconsistency, addressing event-level inconsistency following the inherent subject-predicate logic of news and robustly learning news representations from poor-quality news samples remain two challenges. In this paper, we propose an Event-diven fake news detection framework (Event-Radar) based on multi-view learning, which integrates visual manipulation, textual emotion and multimodal inconsistency at event-level for fake news detection. Specifically, leveraging the capability of graph structures to capture interactions between events and parameters, Event-Radar captures event-level multimodal inconsistency by constructing an event graph that includes multimodal entity subject-predicate logic. Additionally, to mitigate the interference of poor-quality news, Event-Radar introduces a multi-view fusion mechanism, learning comprehensive and robust representations by computing the credibility of each view as a clue, thereby detecting fake news. Extensive experiments demonstrate that Event-Radar achieves outstanding performance on three large-scale fake news detection benchmarks. Our studies also confirm that Event-Radar exhibits strong robustness, providing a paradigm for detecting fake news from noisy news samples.

Deep neural networks exhibit vulnerability to word-level adversarial attacks in natural language processing. Most of these attack methods adopt synonymous substitutions to perturb original samples for crafting adversarial examples while attempting to maintain semantic consistency with the originals. Some of them claim that they could achieve over 90% attack success rate, thereby raising serious safety concerns. However, our investigation reveals that many purportedly successful adversarial examples are actually invalid due to significant changes in semantic meanings compared to their originals. Even when equipped with semantic constraints such as BERTScore, existing attack methods can generate up to 87.9% invalid adversarial examples. Building on this insight, we first curate a 13K dataset for adversarial validity evaluation with the help of GPT-4. Then, an open-source large language model is fine-tuned to offer an interpretable validity score for assessing the semantic consistency between original and adversarial examples. Finally, this validity score can serve as a guide for existing adversarial attack methods to generate valid adversarial examples. Comprehensive experiments demonstrate the effectiveness of our method in evaluating and refining the quality of adversarial examples.

Controlling the attribute intensity of text generation is crucial across scenarios (e.g., writing conciseness, chatting emotion, and explanation clarity). The remarkable capabilities of large language models (LLMs) have revolutionized text generation, prompting us to explore such smooth control of LLM generation. Specifically, we propose metrics to assess the range, calibration, and consistency of the generated text’s attribute intensity in response to varying control values, as well as its relevance to the intended context. To quantify the attribute intensity and context relevance, we leverage an Elo rating system and GPT4, respectively, both renowned for their robust alignment with human judgment. We look into two viable training-free methods for achieving smooth control of LLMs: (1) Prompting with semantic shifters, and (2) Modifying internal model representations. The evaluations of these two methods are conducted on 5 different attributes with various models.

Large language models (LLMs) need to serve everyone, including a global majority of non-English speakers. However, most LLMs today, and open LLMs in particular, are often intended for use in just English (e.g. Llama2, Mistral) or a small handful of high-resource languages (e.g. Mixtral, Qwen). Recent research shows that, despite limits in their intended use, people prompt LLMs in many different languages.Therefore, in this paper, we investigate the basic multilingual capabilities of state-of-the-art open LLMs beyond their intended use.For this purpose, we introduce MultiQ, a new silver standard benchmark for basic open-ended question answering with 27.4k test questions across a typologically diverse set of 137 languages. With MultiQ, we evaluate language fidelity, i.e. whether models respond in the prompted language, and question answering accuracy. All LLMs we test respond faithfully and/or accurately for at least some languages beyond their intended use. Most models are more accurate when they respond faithfully. However, differences across models are large, and there is a long tail of languages where models are neither accurate nor faithful. We explore differences in tokenization as a potential explanation for our findings, identifying possible correlations that warrant further investigation.

Existing works on long-term open-domain dialogues focus on evaluating model responses within contexts spanning no more than five chat sessions. Despite advancements in long-context large language models (LLMs) and retrieval augmented generation (RAG) techniques, their efficacy in very long-term dialogues remains unexplored. To address this research gap, we introduce a machine-human pipeline to generate high-quality, very long-term dialogues by leveraging LLM-based agent architectures and grounding their dialogues on personas and temporal event graphs. Moreover, we equip each agent with the capability of sharing and reacting to images. The generated conversations are verified and edited by human annotators for long-range consistency and grounding to the event graphs. Using this pipeline, we collect LoCoMo, a dataset of very long-term conversations, each encompassing 600 turns and 16K tokens on avg., over up to 32 sessions. Based on LoCoMo, we present a comprehensive evaluation benchmark to measure long-term memory in models, encompassing question answering, event summarization, and multi-modal dialogue generation tasks. Our experimental results indicate that LLMs exhibit challenges in understanding lengthy conversations and comprehending long-range temporal and causal dynamics within dialogues. Employing strategies like long-context LLMs or RAG can offer improvements but these models still substantially lag behind human performance.

Compositional generalization refers to the ability to generalize to novel combinations of previously observed words and syntactic structures.Since it is regarded as a desired property of neural models, recent work has assessed compositional generalization in machine translation as well as semantic parsing.However, previous evaluations with machine translation have focused mostly on lexical generalization (i.e., generalization to unseen combinations of known words).Thus, it remains unclear to what extent models can translate sentences that require structural generalization (i.e., generalization to different sorts of syntactic structures).To address this question, we construct SGET, a machine translation dataset covering various types of compositional generalization with control of words and sentence structures.We evaluate neural machine translation models on SGET and show that they struggle more in structural generalization than in lexical generalization.We also find different performance trends in semantic parsing and machine translation, which indicates the importance of evaluations across various tasks.

Generative search engines have the potential to transform how people seek information online, but generated responses from existing large language models (LLMs)-backed generative search engines may not always be accurate. Nonetheless, retrieval-augmented generation exacerbates safety concerns, since adversaries may successfully evade the entire system by subtly manipulating the most vulnerable part of a claim. To this end, we propose evaluating the robustness of generative search engines in the realistic and high-risk setting, where adversaries have only black-box system access and seek to deceive the model into returning incorrect responses. Through a comprehensive human evaluation of various generative search engines, such as Bing Chat, PerplexityAI, and YouChat across diverse queries, we demonstrate the effectiveness of adversarial factual questions in inducing incorrect responses. Moreover, retrieval-augmented generation exhibits a higher susceptibility to factual errors compared to LLMs without retrieval. These findings highlight the potential security risks of these systems and emphasize the need for rigorous evaluation before deployment. The dataset and code will be publicly available.

The rapid advancement of Large Language Models (LLMs) in the realm of mathematical reasoning necessitates comprehensive evaluations to gauge progress and inspire future directions. Existing assessments predominantly focus on problem-solving from the examinee perspective, overlooking a dual perspective of examiner regarding error identification and correction.From the examiner perspective, we define four evaluation tasks for error identification and correction along with a new dataset with annotated error types and steps. We also design diverse prompts to thoroughly evaluate eleven representative LLMs. Our principal findings indicate that GPT-4 outperforms all models, while open-source model LLaMA-2-7B demonstrates comparable abilities to closed-source models GPT-3.5 and Gemini Pro.Notably, calculation error proves the most challenging error type. Moreover, prompting LLMs with the error types can improve the average correction accuracy by 47.9%. These results reveal potential directions for developing the mathematical reasoning abilities of LLMs.Our code and dataset is available on https://github.com/LittleCirc1e/EIC.

Educational materials such as survey articles in specialized fields like computer science traditionally require tremendous expert inputs and are therefore expensive to create and update. Recently, Large Language Models (LLMs) have achieved significant success across various general tasks. However, their effectiveness and limitations in the education domain are yet to be fully explored. In this work, we examine the proficiency of LLMs in generating succinct survey articles specific to the niche field of NLP in computer science, focusing on a curated list of 99 topics. Automated benchmarks reveal that GPT-4 surpasses its predecessors, inluding GPT-3.5, PaLM2, and LLaMa2 by margins ranging from 2% to 20% in comparison to the established ground truth. We compare both human and GPT-based evaluation scores and provide in-depth analysis. While our findings suggest that GPT-created surveys are more contemporary and accessible than human-authored ones, certain limitations were observed. Notably, GPT-4, despite often delivering outstanding content, occasionally exhibited lapses like missing details or factual errors. At last, we compared the rating behavior between humans and GPT-4 and found systematic bias in using GPT evaluation.

As corporations rush to integrate large language models (LLMs) it is critical that they provide factually accurate information, that is robust to any presuppositions that a user may express. In this work, we introduce UPHILL, a dataset consisting of health-related queries with varying degrees of presuppositions. Using UPHILL, we evaluate the factual accuracy and consistency of InstructGPT, ChatGPT, GPT-4 and Bing Copilot models. We find that while model responses rarely contradict true health claims (posed as questions), all investigated models fail to challenge false claims. Alarmingly, responses from these models agree with 23-32% of the existing false claims, and 49-55% with novel fabricated claims. As we increase the extent of presupposition in input queries, responses from all models except Bing Copilot agree with the claim considerably more often, regardless of its veracity. Given the moderate factual accuracy, and the inability of models to challenge false assumptions, our work calls for a careful assessment of current LLMs for use in high-stakes scenarios.