While recent research increasingly emphasizes the value of human-LLM collaboration in competitive programming and proposes numerous empirical methods, a comprehensive understanding remains elusive due to the fragmented nature of existing studies and their use of diverse, application-specific human feedback. Thus, our work serves a three-fold purpose: First, we present the first taxonomy of human feedback consolidating the entire programming process, which promotes fine-grained evaluation. Second, we introduce ELABORATIONSET, a novel programming dataset specifically designed for human-LLM collaboration, meticulously annotated to enable large-scale simulated human feedback and facilitate cost-effective real human interaction studies. Third, we introduce ELABORATION, a novel benchmark to facilitate a thorough assessment of human-LLM competitive programming. With ELABORATION, we pinpoint strengthes and weaknesses of existing methods, thereby setting the foundation for furture improvement. Our dataset and code will be openly released.

Chain-of-thought (CoT) significantly enhances the performance of large language models (LLMs) across a wide range of tasks, and prior research shows that CoT can theoretically increase expressiveness. However, there is limited mechanistic understanding of the algorithms that Transformer+CoT can learn. Our key contributions are: (1) We evaluate the state tracking capabilities of Transformer+CoT and its variants, confirming the effectiveness of CoT. (2) Next, we identify the circuit (a subset of model components, responsible for tracking the world state), indicating that late-layer MLP neurons play a key role. We propose two metrics, compression and distinction, and show that the neuron sets for each state achieve nearly 100% accuracy, providing evidence of an implicit finite state automaton (FSA) embedded within the model. (3) Additionally, we explore three challenging settings: skipping intermediate steps, introducing data noises, and testing length generalization. Our results demonstrate that Transformer+CoT learns robust algorithms (FSAs), highlighting its resilience in challenging scenarios. Our code is available at https://github.com/IvanChangPKU/FSA.

This work introduces RARE (Retrieval-Augmented Reasoning Enhancement), a versatile extension to the mutual reasoning framework (rStar), aimed at enhancing reasoning accuracy and factual integrity across large language models (LLMs) for complex, knowledge-intensive tasks such as medical and commonsense reasoning. RARE incorporates two innovative actions within the Monte Carlo Tree Search (MCTS) framework: (A6), which generates search queries based on the initial problem statement, performs information retrieval using those queries, and augments reasoning with the retrieved data to formulate the final answer; and (A7), which leverages information retrieval specifically for generated sub-questions and re-answers these sub-questions with the relevant contextual information. Additionally, a Retrieval-Augmented Factuality Scorer is proposed to replace the original discriminator, prioritizing reasoning paths that meet high standards of factuality. Experimental results with LLaMA 3.1 show that RARE enables open-source LLMs to achieve competitive performance with top closed-source models like GPT-4 and GPT-4o. This research establishes RARE as a scalable solution for improving LLMs in domains where logical coherence and factual integrity are critical.

Visual mathematical reasoning, as a fundamental visual reasoning ability, has received widespread attention from the Large Multimodal Models (LMMs) community. Existing benchmarks mainly focus more on the end-to-end performance, but neglect the underlying principles of knowledge acquisition and generalization. Instead, we introduce WE-MATH, the first benchmark specifically designed to explore the problem-solving principles. We meticulously collect 6.5K visual math problems and decompose them into 10.9K step-level questions for evaluation, spanning 5 layers of knowledge granularity and 67 hierarchical knowledge concepts. Specifically, we decompose composite problems into sub-problems according to the required knowledge concepts and introduce a novel four-dimensional metric to hierarchically assess inherent issues in LMMs’ reasoning process. With WE-MATH, we conduct a thorough evaluation of existing LMMs in visual mathematical reasoning and provide comprehensive analysis and insight for future development. We anticipate that WE-MATH will open new pathways for advancements in visual mathematical reasoning for LMMs. Data and code are available at https://github.com/We-Math/We-Math.

Oracle Bone Script (OBS) is a vital treasure of human civilization, rich in insights from ancient societies. However, the evolution of written language over millennia complicates its decipherment. In this paper, we propose V-Oracle, an innovative framework that utilizes Large Multi-modal Models (LMMs) for interpreting OBS. V-Oracle applies principles of pictographic character formation and frames the task as a visual question-answering (VQA) problem, establishing a multi-step reasoning chain. It proposes a multi-dimensional data augmentation for synthesizing high-quality OBS samples, and also implements a multi-phase oracle alignment tuning to improve LMMs’ visual reasoning capabilities. Moreover, to bridge the evaluation gap in the OBS field, we further introduce Oracle-Bench, a comprehensive benchmark that emphasizes process-oriented assessment and incorporates both standard and out-of-distribution setups for realistic evaluation. Extensive experimental results can demonstrate the effectiveness of our method in providing quantitative analyses and superior deciphering capability.

We present Autonomous Data Selection (AutoDS), a method that leverages base language models as zero-shot “generative classifiers” to automatically curate high-quality mathematical texts. Unlike prior approaches that require human annotations or training a dedicated data filter, AutoDS relies solely on a model’s logits to determine whether a given passage is mathematically informative and educational. By integrating AutoDS into a continual pretraining pipeline, we substantially boost downstream performance on challenging math benchmarks (MATH, GSM8K, and BBH) while using far fewer tokens than previous methods. Empirically, our approach achieves roughly a twofold improvement in pretraining token efficiency over strong baselines, underscoring the potential of self-directed data selection in enhancing mathematical reasoning. We will release our curated dataset to facilitate future research in automated domain-specific data curation.

Automating planning with LLMs presents transformative opportunities for traditional industries, yet remains underexplored. In commercial construction, the complexity of automated scheduling often requires manual intervention to ensure precision. We propose CONSTRUCTA, a novel framework leveraging LLMs to optimize construction schedules in complex projects like semiconductor fabrication. CONSTRUCTA addresses key challenges by: (1) integrating construction-specific knowledge through static RAG; (2) employing context-sampling techniques inspired by architectural expertise to provide relevant input; and (3) deploying Construction DPO to align schedules with expert preferences using RLHF. Experiments on proprietary data demonstrate performance improvements of +42.3% in missing value prediction, +79.1% in dependency analysis, and +28.9% in automated planning compared to baseline methods, showcasing its potential to revolutionize construction workflows and inspire domain-specific LLM advancements.

The performance of Large Language Models (LLMs) is substantially influenced by the pretraining corpus, which consists of vast quantities of unsupervised data processed by the models. Despite its critical role in model performance, ensuring the quality of this data is challenging due to its sheer volume and the absence of sample-level quality annotations and enhancements. In this paper, we introduce DecorateLM, a data engineering method designed to refine the pretraining corpus through data rating, tagging and editing. Specifically, DecorateLM rates texts against quality criteria, tags texts with hierarchical labels, and edits texts into a more formalized format. Due to the massive size of the pretraining corpus, adopting an LLM for decorating the entire corpus is less efficient. Therefore, to balance performance with efficiency, we curate a meticulously annotated training corpus for DecorateLM using a large language model and distill data engineering expertise into a compact 1.2 billion parameter small language model (SLM). We then apply DecorateLM to enhance 100 billion tokens of the training corpus, selecting 45 billion tokens that exemplify high quality and diversity for the further training of another 1.2 billion parameter LLM. Our results demonstrate that employing such high-quality data can significantly boost model performance, showcasing a powerful approach to enhance the quality of the pretraining corpus.

Pre-trained language models (LMs) have shown remarkable reasoning performance using explanations or chain-of-thoughts (CoT)) for in-context learning. On the other hand, these reasoning tasks are usually presumed to be more approachable for symbolic programming. To understand the mechanism of reasoning of LMs, we curate synthetic datasets containing equivalent (natural, symbolic) data pairs, where symbolic examples contain first-order logic rules and predicates from non-parametric knowledge bases (KBs), supporting automated verification of intermediate reasoning results. Then we revisit neuro-symbolic approaches and propose to learn from demonstrations containing logic rules and corresponding examples to iteratively reason over KBs, recovering Prolog’s backward chaining algorithm and supporting automated verification of LMs’ outputs. Comprehensive experiments are included to systematically compare LMLP with CoT in deductive reasoning settings, showing that LMLP enjoys more than 25% higher accuracy than CoT on length generalization benchmarks even with smaller model sizes.

Multi-modal Event Reasoning (MMER) endeavors to endow machines with the ability to comprehend intricate event relations across diverse data modalities. MMER is fundamental and underlies a wide broad of applications. Despite extensive instruction fine-tuning, current multi-modal large language models still fall short in such ability. The disparity stems from that existing models are insufficient to capture underlying principles governing event evolution in various scenarios. In this paper, we introduce Multi-Modal Event Evolution Learning (MEEL) to enable the model to grasp the event evolution mechanism yielding advanced MMER ability. Specifically, we commence with the design of event diversification to gather seed events from a rich spectrum of scenarios. Subsequently, we employ ChatGPT to generate evolving graphs for these seed events. We propose an instruction encapsulation process that formulates the evolving graphs into instruction-tuning data, aligning the comprehension of event reasoning to humans. Finally, we observe that models trained in this way are still struggling to fully comprehend event evolution. In such a case, we propose the guiding discrimination strategy, in which models are trained to discriminate the improper evolution direction. We collect and curate a benchmark M-EV2 for MMER. Extensive experiments on M-EV2 validate the effectiveness of our approach, showcasing competitive performance in open-source multi-modal LLMs.

Accurate uncertainty quantification is crucial for the safe deployment of machine learning models, and prior research has demonstrated improvements in the calibration of modern language models (LMs). We study in-context learning (ICL), a prevalent method for adapting static LMs through tailored prompts, and examine the balance between performance and calibration across a broad spectrum of natural language understanding and reasoning tasks. Through comprehensive experiments, we observe that, with an increasing number of ICL examples, models initially exhibit increased miscalibration before achieving better calibration and miscalibration tends to arise in low-shot settings. Moreover, we find that methods aimed at improving usability, such as fine-tuning and chain-of-thought (CoT) prompting, can lead to miscalibration and unreliable natural language explanations. Furthermore, we explore recalibration techniques and find that a scaling-binning calibrator can reduce calibration errors consistently.

With the emergence of the COVID-19 pandemic, the political and the medical aspects of disinformation merged as the problem got elevated to a whole new level to become the first global infodemic. Fighting this infodemic is currently ranked very high on the list of priorities of the World Health Organization, with dangers ranging from promoting fake cures, rumors, and conspiracy theories to spreading xenophobia and panic. With this in mind, we studied how COVID-19 is discussed in Bulgarian social media in terms of factuality, harmfulness, propaganda, and framing. We found that most Bulgarian tweets contain verifiable factual claims, are factually true, are of potential public interest, are not harmful, and are too trivial to fact-check; moreover, zooming into harmful tweets, we found that they spread not only rumors but also panic. We further analyzed articles shared in Bulgarian partisan pro/con-COVID-19 Facebook groups and found that propaganda is more prevalent in skeptical articles, which use doubt, flag waving, and slogans to convey their message; in contrast, concerned ones appeal to emotions, fear, and authority; moreover, skeptical articles frame the issue as one of quality of life, policy, legality, economy, and politics, while concerned articles focus on health & safety. We release our manually and automatically analyzed datasets to enable further research.

While COVID-19 vaccines are finally becoming widely available, a second pandemic that revolves around the circulation of anti-vaxxer “fake news” may hinder efforts to recover from the first one. With this in mind, we performed an extensive analysis of Arabic and English tweets about COVID-19 vaccines, with focus on messages originating from Qatar. We found that Arabic tweets contain a lot of false information and rumors, while English tweets are mostly factual. However, English tweets are much more propagandistic than Arabic ones. In terms of propaganda techniques, about half of the Arabic tweets express doubt, and 1/5 use loaded language, while English tweets are abundant in loaded language, exaggeration, fear, name-calling, doubt, and flag-waving. Finally, in terms of framing, Arabic tweets adopt a health and safety perspective, while in English economic concerns dominate.

Opinion prediction is an emerging research area with diverse real-world applications, such as market research and situational awareness. We identify two lines of approaches to the problem of opinion prediction. One uses topic-based sentiment analysis with time-series modeling, while the other uses static embedding of text. The latter approaches seek user-specific solutions by generating user fingerprints. Such approaches are useful in predicting user’s reactions to unseen content. In this work, we propose a novel dynamic fingerprinting method that leverages contextual embedding of user’s comments conditioned on relevant user’s reading history. We integrate BERT variants with a recurrent neural network to generate predictions. The results show up to 13% improvement in micro F1-score compared to previous approaches. Experimental results show novel insights that were previously unknown such as better predictions for an increase in dynamic history length, the impact of the nature of the article on performance, thereby laying the foundation for further research.

Recent events, such as the 2016 US Presidential Campaign, Brexit and the COVID-19 “infodemic”, have brought into the spotlight the dangers of online disinformation. There has been a lot of research focusing on fact-checking and disinformation detection. However, little attention has been paid to the specific rhetorical and psychological techniques used to convey propaganda messages. Revealing the use of such techniques can help promote media literacy and critical thinking, and eventually contribute to limiting the impact of “fake news” and disinformation campaigns. Prta (Propaganda Persuasion Techniques Analyzer) allows users to explore the articles crawled on a regular basis by highlighting the spans in which propaganda techniques occur and to compare them on the basis of their use of propaganda techniques. The system further reports statistics about the use of such techniques, overall and over time, or according to filtering criteria specified by the user based on time interval, keywords, and/or political orientation of the media. Moreover, it allows users to analyze any text or URL through a dedicated interface or via an API. The system is available online: https://www.tanbih.org/prta.

We introduce Tanbih, a news aggregator with intelligent analysis tools to help readers understanding what’s behind a news story. Our system displays news grouped into events and generates media profiles that show the general factuality of reporting, the degree of propagandistic content, hyper-partisanship, leading political ideology, general frame of reporting, and stance with respect to various claims and topics of a news outlet. In addition, we automatically analyse each article to detect whether it is propagandistic and to determine its stance with respect to a number of controversial topics.

This paper presents QCRI’s Arabic-to-English live speech translation system. It features modern web technologies to capture live audio, and broadcasts Arabic transcriptions and English translations simultaneously. Our Kaldi-based ASR system uses the Time Delay Neural Network (TDNN) architecture, while our Machine Translation (MT) system uses both phrase-based and neural frameworks. Although our neural MT system is slower than the phrase-based system, it produces significantly better translations and is memory efficient. The demo is available at https://st.qcri.org/demos/livetranslation.

Sentiment classification becomes more and more important with the rapid growth of user generated content. However, sentiment classification task usually comes with two challenges: first, sentiment classification is highly domain-dependent and training sentiment classifier for every domain is inefficient and often impractical; second, since the quantity of labeled data is important for assessing the quality of classifier, it is hard to evaluate classifiers when labeled data is limited for certain domains. To address the challenges mentioned above, we focus on learning high-level features that are able to generalize across domains, so a global classifier can benefit with a simple combination of documents from multiple domains. In this paper, the proposed model incorporates both sentiment polarity and unlabeled data from multiple domains and learns new feature representations. Our model doesn’t require labels from every domain, which means the learned feature representation can be generalized for sentiment domain adaptation. In addition, the learned feature representation can be used as classifier since our model defines the meaning of feature value and arranges high-level features in a prefixed order, so it is not necessary to train another classifier on top of the new features. Empirical evaluations demonstrate our model outperforms baselines and yields competitive results to other state-of-the-art works on benchmark datasets.