Direct Preference Optimization (DPO) optimizes language models to align with human preferences. Utilizing on-policy samples, generated directly by the policy model, typically results in better performance due to its distribution consistency with the model compared to off-policy samples. This paper identifies the quality of candidate preference samples as another critical factor. While the quality of on-policy data is inherently constrained by the capabilities of the policy model, off-policy data, which can be derived from diverse sources, offers greater potential for quality despite experiencing distribution shifts. However, current research mostly relies on on-policy data and neglects the value of off-policy data in terms of data quality, due to the challenge posed by distribution shift. In this paper, we propose InCo-DPO, an efficient method for synthesizing preference data by integrating on-policy and off-policy data, allowing dynamic adjustments to balance distribution shifts and data quality, thus finding an optimal trade-off. Consequently, InCo-DPO overcomes the limitations of distribution shifts in off-policy data and the quality constraints of on-policy data. We evaluated InCo-DPO with the Alpaca-Eval 2.0 and Arena-Hard benchmarks. Experimental results demonstrate that our approach not only outperforms both on-policy and off-policy data but also achieves a state-of-the-art win rate of 60.8 on Arena-Hard with the vanilla DPO using Gemma-2 model.

Recently, Vision-Language Models (VLMs) have achieved remarkable progress in multimodal tasks, and multimodal instruction data serves as the foundation for enhancing VLM capabilities. Despite the availability of several open-source multimodal datasets, limitations in the scale and quality of open-source instruction data hinder the performance of VLMs trained on these datasets, leading to a significant gap compared to models trained on closed-source data. To address this challenge, we introduce Infinity-MM, a large-scale multimodal instruction dataset. We collected the available multimodal instruction datasets and performed unified preprocessing, resulting in a dataset with over 40 million samples that ensures diversity and accuracy. Furthermore, to enable large-scale expansion of instruction data and support the continuous acquisition of high-quality data, we propose a synthetic instruction generation method based on a tagging system and open-source VLMs. By establishing correspondences between different types of images and associated instruction types, this method can provide essential guidance during data synthesis. Leveraging this high-quality data, we have trained a 2-billion-parameter Vision-Language Model, Aquila-VL-2B, which achieves state-of-the-art (SOTA) performance among models of similar scale. The data is available at: https://huggingface.co/datasets/BAAI/Infinity-MM.

The success of Large Language Models (LLMs) [1][2] is primarily attributed to the availability of extensive, high-quality pre-training corpora, which underpin their foundational knowledge and reasoning capabilities for a variety of tasks, from creative writing to complex problem-solving. Among them, the Open-source datasets, such as The Pile[3] and Common Crawl[4], have been instrumental in propelling LLM development, fostering collaboration and establishing benchmarks for innovation. Existing Researchers focus more on scaling high-quality data. Recently the demand for pre-training data has exceeded 10 trillion tokens [1][5][6], underscoring two key trajectories in English pre-training: scaling data and improving its quality. Open-source datasets have rapidly expanded, evolving from collections like the Pile(825GB) to larger datasets such as FineWeb(15TB)[7], which draw extensively from Common Crawl. Simultaneously, the focus has shifted from rule-based filtering methods, as seen in early projects like Redpajama[8], to model-driven approaches exemplified by FineWeb-Edu[7]. Despite the rapid advancement of English open-source datasets, Chinese data remains significantly underrepresented on the global web. Existing open-source Chinese datasets, such as WuDao [9], SkyPile150B [10], and WanjuanV1 [11], are constrained in scale due to a scarcity of Chinese data sources online. Furthermore, there is limited research focused on improving quality classification for Chinese web data, resulting in suboptimal data quality.

Tokenizer is an essential component for large language models (LLMs), and a tokenizer with a high compression rate can improve the model's representation and processing efficiency. However, the tokenizer cannot ensure high compression rate in all scenarios, and an increase in the average input and output lengths will increases the training and inference costs of the model. Therefore, it is crucial to find ways to improve the model's efficiency with minimal cost while maintaining the model's performance. In this work, we propose a method to improve model representation and processing efficiency by replacing the tokenizers of LLMs. We propose replacing and reinitializing the parameters of the model's input and output layers with the parameters of the original model, and training these parameters while keeping other parameters fixed. We conducted experiments on different LLMs, and the results show that our method can maintain the performance of the model after replacing the tokenizer, while significantly improving the decoding speed for long texts.

The goal of dialogue state tracking (DST) is to predict the current dialogue state given all previous dialogue contexts. Existing approaches generally predict the dialogue state at every turn from scratch. However, the overwhelming majority of the slots in each turn should simply inherit the slot values from the previous turn. Therefore, the mechanism of treating slots equally in each turn not only is inefficient but also may lead to additional errors because of the redundant slot value generation. To address this problem, we devise the two-stage DSS-DST which consists of the Dual Slot Selector based on the current turn dialogue, and the Slot Value Generator based on the dialogue history. The Dual Slot Selector determines each slot whether to update slot value or to inherit the slot value from the previous turn from two aspects: (1) if there is a strong relationship between it and the current turn dialogue utterances; (2) if a slot value with high reliability can be obtained for it through the current turn dialogue. The slots selected to be updated are permitted to enter the Slot Value Generator to update values by a hybrid method, while the other slots directly inherit the values from the previous turn. Empirical results show that our method achieves 56.93%, 60.73%, and 58.04% joint accuracy on MultiWOZ 2.0, MultiWOZ 2.1, and MultiWOZ 2.2 datasets respectively and achieves a new state-of-the-art performance with significant improvements.