SailCompass: Towards Reproducible and Robust Evaluation for Southeast Asian Languages

In this paper, we introduce SailCompass, a reproducible and robust evaluation benchmark for assessing Large Language Models (LLMs) on Southeast Asian Languages (SEA). SailCompass encompasses three main SEA languages, eight primary tasks including 14 datasets covering three task types (generation, multiple-choice questions, and classification). To improve the robustness of the evaluation approach, we explore different prompt configurations for multiple-choice questions and leverage calibrations to improve the faithfulness of classification tasks. With SailCompass, we derive the following findings: (1) SEA-specialized LLMs still outperform general LLMs, although the gap has narrowed; (2) A balanced language distribution is important for developing better SEA-specialized LLMs; (3) Advanced prompting techniques (e.g., calibration, perplexity-based ranking) are necessary to better utilize LLMs. All datasets and evaluation scripts are public.

When Precision Meets Position: BFloat16 Breaks Down RoPE in Long-Context Training

Extending context window sizes allows large language models (LLMs) to process longer sequences and handle more complex tasks. Rotary Positional Embedding (RoPE) has become the de facto standard due to its relative positional encoding properties that benefit long-context training. However, we observe that using RoPE with BFloat16 format results in numerical issues, causing it to deviate from its intended relative positional encoding, especially in long-context scenarios. This issue arises from BFloat16's limited precision and accumulates as context length increases, with the first token contributing significantly to this problem. To address this, we develop AnchorAttention, a plug-and-play attention method that alleviates numerical issues caused by BFloat16, improves long-context capabilities, and speeds up training. AnchorAttention reduces unnecessary attention computations, maintains semantic coherence, and boosts computational efficiency by treating the first token as a shared anchor with a consistent position ID, making it visible to all documents within the training context. Experiments on three types of LLMs demonstrate that AnchorAttention significantly improves long-context performance and reduces training time by over 50\% compared to standard full attention mechanisms, while preserving the original LLM's capabilities on general tasks. Our code is available at https://github.com/haonan3/AnchorContext.

Spider 2.0: Evaluating Language Models on Real-World Enterprise Text-to-SQL Workflows

Real-world enterprise text-to-SQL workflows often involve complex cloud or local data across various database systems, multiple SQL queries in various dialects, and diverse operations from data transformation to analytics. We introduce Spider 2.0, an evaluation framework comprising 632 real-world text-to-SQL workflow problems derived from enterprise-level database use cases. The databases in Spider 2.0 are sourced from real data applications, often containing over 1,000 columns and stored in local or cloud database systems such as BigQuery and Snowflake. We show that solving problems in Spider 2.0 frequently requires understanding and searching through database metadata, dialect documentation, and even project-level codebases. This challenge calls for models to interact with complex SQL workflow environments, process extremely long contexts, perform intricate reasoning, and generate multiple SQL queries with diverse operations, often exceeding 100 lines, which goes far beyond traditional text-to-SQL challenges. Our evaluations indicate that based on o1-preview, our code agent framework successfully solves only 17.0% of the tasks, compared with 91.2% on Spider 1.0 and 73.0% on BIRD. Our results on Spider 2.0 show that while language models have demonstrated remarkable performance in code generation -- especially in prior text-to-SQL benchmarks -- they require significant improvement in order to achieve adequate performance for real-world enterprise usage. Progress on Spider 2.0 represents crucial steps towards developing intelligent, autonomous, code agents for real-world enterprise settings. Our code, baseline models, and data are available at https://spider2-sql.github.io.

OpenCoder: The Open Cookbook for Top-Tier Code Large Language Models

Large language models (LLMs) for code have become indispensable in various domains, including code generation, reasoning tasks and agent systems.While open-access code LLMs are increasingly approaching the performance levels of proprietary models, high-quality code LLMs suitable for rigorous scientific investigation, particularly those with reproducible data processing pipelines and transparent training protocols, remain limited. The scarcity is due to various challenges, including resource constraints, ethical considerations, and the competitive advantages of keeping models advanced. To address the gap, we introduce OpenCoder, a top-tier code LLM that not only achieves performance comparable to leading models but also serves as an ``open cookbook'' for the research community. Unlike most prior efforts, we release not only model weights and inference code, but also the reproducible training data, complete data processing pipeline, rigorous experimental ablation results, and detailed training protocols for open scientific research. Through this comprehensive release, we identify the key ingredients for building a top-tier code LLM: (1) code optimized heuristic rules for data cleaning and methods for data deduplication, (2) recall of text corpus related to code and (3) high-quality synthetic data in both annealing and supervised fine-tuning stages. By offering this level of openness, we aim to broaden access to all aspects of a top-tier code LLM, with OpenCoder serving as both a powerful model and an open foundation to accelerate research, and enable reproducible advancements in code AI.

Scaling up Masked Diffusion Models on Text

Masked diffusion models (MDMs) have shown promise in language modeling, yet their scalability and effectiveness in core language tasks, such as text generation and language understanding, remain underexplored. This paper establishes the first scaling law for MDMs, demonstrating a scaling rate comparable to autoregressive models (ARMs) and a relatively small compute gap. Motivated by their scalability, we train a family of MDMs with up to 1.1 billion (B) parameters to systematically evaluate their performance against ARMs of comparable or larger sizes. Fully leveraging the probabilistic formulation of MDMs, we propose a simple yet effective \emph{unsupervised classifier-free guidance} that effectively exploits large-scale unpaired data, boosting performance for conditional inference. In language understanding, a 1.1B MDM shows competitive results, outperforming the larger 1.5B GPT-2 model on four out of eight zero-shot benchmarks. In text generation, MDMs provide a flexible trade-off compared to ARMs utilizing KV-cache: MDMs match the performance of ARMs while being 1.4 times faster, or achieve higher quality than ARMs at a higher computational cost. Moreover, MDMs address challenging tasks for ARMs by effectively handling bidirectional reasoning and adapting to temporal shifts in data. Notably, a 1.1B MDM breaks the \emph{reverse curse} encountered by much larger ARMs with significantly more data and computation, such as Llama-2 (13B) and GPT-3 (175B). Our code is available at \url{https://github.com/ML-GSAI/SMDM}.

DataTales: A Benchmark for Real-World Intelligent Data Narration

We introduce DataTales, a novel benchmark designed to assess the proficiency of language models in data narration, a task crucial for transforming complex tabular data into accessible narratives. Existing benchmarks often fall short in capturing the requisite analytical complexity for practical applications. DataTales addresses this gap by offering 4.9k financial reports paired with corresponding market data, showcasing the demand for models to create clear narratives and analyze large datasets while understanding specialized terminology in the field. Our findings highlights the significant challenge that language models face in achieving the necessary precision and analytical depth for proficient data narration, suggesting promising avenues for future model development and evaluation methodologies.

Joint Space-Time Adaptive Processing and Beamforming Design for Cell-Free ISAC Systems

In this paper, we explore cooperative sensing and communication within cell-free integrated sensing and communication (ISAC) systems. Specifically, multiple transmit access points (APs) collaboratively serve multiple communication users while simultaneously illuminating a potential target, with a separate sensing AP dedicated to collecting echo signals for target detection. To improve the performance of identifying a moving target in the presence of strong interference originating from transmit APs, we employ the space-time adaptive processing (STAP) technique and jointly optimize the transmit/receive beamforming. Our goal is to maximize the radar output signal-to-interference-plus-noise ratio (SINR), subject to the communication SINR requirements and the power budget. An efficient alternating algorithm is developed to solve the resulting non-convex optimization problem. Simulations demonstrate significant performance improvements in target detection and validate the advantages of the proposed joint STAP and beamforming design for cell-free ISAC systems.

DOA Estimation-Oriented Joint Array Partitioning and Beamforming Designs for ISAC Systems

Integrated sensing and communication has been identified as an enabling technology for forthcoming wireless networks. In an effort to achieve an improved performance trade-off between multiuser communications and radar sensing, this paper considers a dynamically-partitioned antenna array architecture for monostatic ISAC systems, in which each element of the array at the base station can function as either a transmit or receive antenna. To fully exploit the available spatial degrees of freedom for both communication and sensing functions, we jointly design the partitioning of the array between transmit and receive antennas together with the transmit beamforming in order to minimize the direction-of-arrival (DOA) estimation error, while satisfying constraints on the communication signal-to-interference-plus-noise ratio and the transmit power budget. An alternating algorithm based on Dinkelbach's transform, the alternative direction method of multipliers, and majorization-minimization is developed to solve the resulting complicated optimization problem. To reduce the computational complexity, we also present a heuristic three-step strategy that optimizes the transmit beamforming after determining the antenna partitioning. Simulation results confirm the effectiveness of the proposed algorithms in significantly reducing the DOA estimation error.

CCSBench: Evaluating Compositional Controllability in LLMs for Scientific Document Summarization

To broaden the dissemination of scientific knowledge to diverse audiences, scientific document summarization must simultaneously control multiple attributes such as length and empirical focus. However, existing research typically focuses on controlling single attributes, leaving the compositional control of multiple attributes underexplored. To address this gap, we introduce CCSBench, a benchmark for compositional controllable summarization in the scientific domain. Our benchmark enables fine-grained control over both explicit attributes (e.g., length), which are objective and straightforward, and implicit attributes (e.g., empirical focus), which are more subjective and conceptual. We conduct extensive experiments on GPT-4, LLaMA2, and other popular LLMs under various settings. Our findings reveal significant limitations in large language models' ability to balance trade-offs between control attributes, especially implicit ones that require deeper understanding and abstract reasoning.

When Attention Sink Emerges in Language Models: An Empirical View

Language Models (LMs) assign significant attention to the first token, even if it is not semantically important, which is known as attention sink. This phenomenon has been widely adopted in applications such as streaming/long context generation, KV cache optimization, inference acceleration, model quantization, and others. Despite its widespread use, a deep understanding of attention sink in LMs is still lacking. In this work, we first demonstrate that attention sinks exist universally in LMs with various inputs, even in small models. Furthermore, attention sink is observed to emerge during the LM pre-training, motivating us to investigate how optimization, data distribution, loss function, and model architecture in LM pre-training influence its emergence. We highlight that attention sink emerges after effective optimization on sufficient training data. The sink position is highly correlated with the loss function and data distribution. Most importantly, we find that attention sink acts more like key biases, storing extra attention scores, which could be non-informative and not contribute to the value computation. We also observe that this phenomenon (at least partially) stems from tokens' inner dependence on attention scores as a result of softmax normalization. After relaxing such dependence by replacing softmax attention with other attention operations, such as sigmoid attention without normalization, attention sinks do not emerge in LMs up to 1B parameters. The code is available at https://github.com/sail-sg/Attention-Sink.