As demonstrated by the proprietary Large Language Models (LLMs) such as GPT and Claude series, LLMs have the potential to achieve remarkable proficiency across a wide range of domains, including law, medicine, finance, science, code, etc., all within a single model. These capabilities are further augmented during the Supervised Fine-Tuning (SFT) phase. Despite their potential, existing work mainly focuses on domain-specific enhancements during fine-tuning, the challenge of which lies in catastrophic forgetting of knowledge across other domains. In this study, we introduce VersaTune, a novel data composition framework designed for enhancing LLMs’ overall multi-domain capabilities during training. We begin with detecting the distribution of domain-specific knowledge within the base model, followed by the training data composition that aligns with the model’s existing knowledge distribution. During the subsequent training process, domain weights are dynamically adjusted based on their learnable potential and forgetting degree. Experimental results indicate that VersaTune is effective in multi-domain fostering, with an improvement of 29.77% in the overall multi-ability performances compared to uniform domain weights. Furthermore, we find that Qwen-2.5-32B + VersaTune even surpasses frontier models, including GPT-4o, Claude3.5-Sonnet and DeepSeek-V3 by 0.86%, 4.76% and 4.60%. Additionally, in scenarios where flexible expansion of a specific domain is required, VersaTune reduces the performance degradation in other domains by 38.77%, while preserving the training efficacy of the target domain.

In recent years, foundation models, particularly large language models (LLMs), have demonstrated significant improvements across a variety of tasks. One of their most important features is long-context capability, which enables them to generate extended text with high semantic coherence, retrieving relevant information, and handling tasks with substantial amounts of text efficiently. The key to improving long-context performance lies in effective data organization and management strategies that integrate data from multiple domains and optimize the context window during training. Through extensive experimental analysis, we identified three key challenges in designing effective data management strategies that enable the model to achieve long-context capability without sacrificing performance in other tasks: (1) a shortage of long documents across multiple domains, (2) effective construction of context windows, and (3) efficient organization of large-scale datasets. To address these challenges, we introduce DataSculpt, a novel data management framework designed for long-context training. We first formulate the organization of training data as a multi-objective optimization problem, focusing on attributes including the relevance among documents within the same training sequence, the quantity of concatenated instances, individual document integrity, and computational cost. Specifically, our approach utilizes a coarse-to-fine method to optimize training data organization effectively. We begin by clustering the data based on semantic similarity (coarse), followed by a multi-objective greedy search within each cluster to score and concatenate documents into various context windows (fine). We have deployed DataSculpt as the data management backend for long-context training in Baichuan Inc. Extensive experiments with diverse downstream tasks show that DataSculpt enhances the model's long-context performance by an average of 15.73%, while maintaining the general capabilities with a 4.63% improvement.

Large Language Models (LLMs) have exhibited significant potential in performing diverse tasks, including the ability to call functions or use external tools to enhance their performance. While current research on function calling by LLMs primarily focuses on single-turn interactions, this paper addresses the overlooked necessity for LLMs to engage in multi-turn function calling—critical for handling compositional, real-world queries that require planning with functions but not only use functions. To facilitate this, we introduce an approach, BUTTON, which generates synthetic compositional instruction tuning data via bottom-up instruction construction and top-down trajectory generation. In the bottom-up phase, we generate simple atomic tasks based on real-world scenarios and build compositional tasks using heuristic strategies based on atomic tasks. Corresponding function definitions are then synthesized for these compositional tasks. The top-down phase features a multi-agent environment where interactions among simulated humans, assistants, and tools are utilized to gather multi-turn function calling trajectories. This approach ensures task compositionality and allows for effective function and trajectory generation by examining atomic tasks within compositional tasks. We produce a dataset BUTTONInstruct comprising 8k data points and demonstrate its effectiveness through extensive experiments across various LLMs.

Books act as a crucial carrier of cultural dissemination in ancient times. This work involves joint efforts between visualization and humanities researchers, aiming at building a holistic view of the cultural exchange and integration between China and Japan brought about by the overseas circulation of Chinese classics. Book circulation data consist of uncertain spatiotemporal trajectories, with multiple dimensions, and movement across hierarchical spaces forms a compound network. LiberRoad visualizes the circulation of books collected in the Imperial Household Agency of Japan, and can be generalized to other book movement data. The LiberRoad system enables a smooth transition between three views (Location Graph, map, and timeline) according to the desired perspectives (spatial or temporal), as well as flexible filtering and selection. The Location Graph is a novel uncertainty-aware visualization method that employs improved circle packing to represent spatial hierarchy. The map view intuitively shows the overall circulation by clustering and allows zooming into single book trajectory with lenses magnifying local movements. The timeline view ranks dynamically in response to user interaction to facilitate the discovery of temporal events. The evaluation and feedback from the expert users demonstrate that LiberRoad is helpful in revealing movement patterns and comparing circulation characteristics of different times and spaces.