Recent studies have been increasingly demonstrating that high-quality data is crucial for effective pretraining of language models. However, the precise definition of "high-quality" remains underexplored. Focusing on the code domain, we introduce Arctic-SnowCoder-1.3B, a data-efficient base code model pretrained on 555B tokens through three phases of progressively refined data: (1) general pretraining with 500B standard-quality code tokens, preprocessed through basic filtering, deduplication, and decontamination, (2) continued pretraining with 50B high-quality tokens, selected from phase one by a BERT-style quality annotator trained to distinguish good code from random data, using positive examples drawn from high-quality code files, along with instruction data from Magicoder and StarCoder2-Instruct, and (3) enhanced pretraining with 5B synthetic data created by Llama-3.1-70B using phase two data as seeds, adapting the Magicoder approach for pretraining. Despite being trained on a limited dataset, Arctic-SnowCoder achieves state-of-the-art performance on BigCodeBench, a coding benchmark focusing on practical and challenging programming tasks, compared to similarly sized models trained on no more than 1T tokens, outperforming Phi-1.5-1.3B by 36%. Across all evaluated benchmarks, Arctic-SnowCoder-1.3B beats StarCoderBase-3B pretrained on 1T tokens. Additionally, it matches the performance of leading small base code models trained on trillions of tokens. For example, Arctic-SnowCoder-1.3B surpasses StarCoder2-3B, pretrained on over 3.3T tokens, on HumanEval+, a benchmark that evaluates function-level code generation, and remains competitive on BigCodeBench. Our evaluation presents a comprehensive analysis justifying various design choices for Arctic-SnowCoder. Most importantly, we find that the key to high-quality data is its alignment with the distribution of downstream applications.

High-resource language models often fall short in the African context, where there is a critical need for models that are efficient, accessible, and locally relevant, even amidst significant computing and data constraints. This paper introduces InkubaLM, a small language model with 0.4 billion parameters, which achieves performance comparable to models with significantly larger parameter counts and more extensive training data on tasks such as machine translation, question-answering, AfriMMLU, and the AfriXnli task. Notably, InkubaLM outperforms many larger models in sentiment analysis and demonstrates remarkable consistency across multiple languages. This work represents a pivotal advancement in challenging the conventional paradigm that effective language models must rely on substantial resources. Our model and datasets are publicly available at https://huggingface.co/lelapa to encourage research and development on low-resource languages.

This paper presents Dolphin, a novel decoder-decoder architecture for energy-efficient processing of long contexts in language models. Our approach addresses the significant energy consumption and latency challenges inherent in on-device models. Dolphin employs a compact 0.5B parameter decoder to distill extensive contextual information into a memory embedding, substantially reducing the input length for the primary 7B parameter decoder model. Inspired by vision-language models, we repurpose the image embedding projector to encode long textual contexts, effectively treating extended context as a distinct modality. This innovative method enables processing of substantially longer contexts without the typical computational overhead associated with extended input sequences. Empirical evaluations demonstrate a 10-fold improvement in energy efficiency and a 5-fold reduction in latency compared to conventional full-length context processing methods without losing quality of the response. Our work contributes to the development of more sustainable and scalable language models for on-device applications, addressing the critical need for energy-efficient and responsive AI technologies in resource-constrained environments while maintaining the accuracy to understand long contexts. This research has implications for the broader field of natural language processing, particularly in the domain of efficient model design for resource-limited settings. By enabling more sophisticated AI capabilities on edge devices, Dolphin paves the way for advanced language processing in a wide range of applications where computational resources are at a premium. The Dolphin model is publicly available at https://huggingface.co/NexaAIDev/Dolphin.

This paper proposes a new method for preventing unsafe or otherwise low quality large language model (LLM) outputs, by leveraging the stochasticity of LLMs. We propose a system whereby LLM checkers vote on the acceptability of a generated output, regenerating it if a threshold of disapproval is reached, until sufficient checkers approve. We further propose estimators for cost and failure rate, and based on those estimators and experimental data tailored to the application, we propose an algorithm that achieves a desired failure rate at the least possible cost. We demonstrate that, under these models, failure rate decreases exponentially as a function of cost when voter count and threshold are chosen according to the algorithm, and that the models reasonably estimate the actual performance of such a system in action, even with limited data.

Building high-quality datasets for specialized tasks is a time-consuming and resource-intensive process that often requires specialized domain knowledge. We propose Corpus Retrieval and Augmentation for Fine-Tuning (CRAFT), a method for generating synthetic datasets, given a small number of user-written few-shots that demonstrate the task to be performed. Given the few-shot examples, we use large-scale public web-crawled corpora and similarity-based document retrieval to find other relevant human-written documents. Lastly, instruction-tuned large language models (LLMs) augment the retrieved documents into custom-formatted task samples, which then can be used for fine-tuning. We demonstrate that CRAFT can efficiently generate large-scale task-specific training datasets for four diverse tasks: biology question-answering (QA), medicine QA and commonsense QA as well as summarization. Our experiments show that CRAFT-based models outperform or achieve comparable performance to general LLMs for QA tasks, while CRAFT-based summarization models outperform models trained on human-curated data by 46 preference points.

We introduce OLMoE, a fully open, state-of-the-art language model leveraging sparse Mixture-of-Experts (MoE). OLMoE-1B-7B has 7 billion (B) parameters but uses only 1B per input token. We pretrain it on 5 trillion tokens and further adapt it to create OLMoE-1B-7B-Instruct. Our models outperform all available models with similar active parameters, even surpassing larger ones like Llama2-13B-Chat and DeepSeekMoE-16B. We present various experiments on MoE training, analyze routing in our model showing high specialization, and open-source all aspects of our work: model weights, training data, code, and logs.

The success of Large Language Models (LLMs) relies heavily on the huge amount of pre-training data learned in the pre-training phase. The opacity of the pre-training process and the training data causes the results of many benchmark tests to become unreliable. If any model has been trained on a benchmark test set, it can seriously hinder the health of the field. In order to automate and efficiently test the capabilities of large language models, numerous mainstream benchmarks adopt a multiple-choice format. As the swapping of the contents of multiple-choice options does not affect the meaning of the question itself, we propose a simple and effective data leakage detection method based on this property. Specifically, we shuffle the contents of the options in the data to generate the corresponding derived data sets, and then detect data leakage based on the model's log probability distribution over the derived data sets. If there is a maximum and outlier in the set of log probabilities, it indicates that the data is leaked. Our method is able to work under black-box conditions without access to model training data or weights, effectively identifying data leakage from benchmark test sets in model pre-training data, including both normal scenarios and complex scenarios where options may have been shuffled intentionally or unintentionally. Through experiments based on two LLMs and benchmark designs, we demonstrate the effectiveness of our method. In addition, we evaluate the degree of data leakage of 31 mainstream open-source LLMs on four benchmark datasets and give a ranking of the leaked LLMs for each benchmark, and we find that the Qwen family of LLMs has the highest degree of data leakage.

This raises the question of whether AI has the potential to shape a fundamental aspect of human culture: the way we speak. Recent analyses revealed that scientific publications already exhibit evidence of AI-specific language. But this evidence is inconclusive, since scientists may simply be using AI to copy-edit their writing. To explore whether AI has influenced human spoken communication, we transcribed and analyzed about 280,000 English-language videos of presentations, talks, and speeches from more than 20,000 YouTube channels of academic institutions. We find a significant shift in the trend of word usage specific to words distinctively associated with ChatGPT following its release. These findings provide the first empirical evidence that humans increasingly imitate LLMs in their spoken language. Our results raise societal and policy-relevant concerns about the potential of AI to unintentionally reduce linguistic diversity, or to be deliberately misused for mass manipulation. They also highlight the need for further investigation into the feedback loops between machine behavior and human culture.

Sequential recommender systems are essential for discerning user preferences from historical interactions and facilitating targeted recommendations. Recent innovations employing Large Language Models (LLMs) have advanced the field by encoding item semantics, yet they often necessitate substantial parameter tuning and are resource-demanding. Moreover, these works fails to consider the diverse characteristics of different types of users and thus diminishes the recommendation accuracy. In this paper, we propose a parameter-efficient Large Language Model Bi-Tuning framework for sequential recommendation with collaborative information (Laser). Specifically, Bi-Tuning works by inserting trainable virtual tokens at both the prefix and suffix of the input sequence and freezing the LLM parameters, thus optimizing the LLM for the sequential recommendation. In our Laser, the prefix is utilized to incorporate user-item collaborative information and adapt the LLM to the recommendation task, while the suffix converts the output embeddings of the LLM from the language space to the recommendation space for the follow-up item recommendation. Furthermore, to capture the characteristics of different types of users when integrating the collaborative information via the prefix, we introduce M-Former, a lightweight MoE-based querying transformer that uses a set of query experts to integrate diverse user-specific collaborative information encoded by frozen ID-based sequential recommender systems, significantly improving the accuracy of recommendations. Extensive experiments on real-world datasets demonstrate that Laser can parameter-efficiently adapt LLMs to effective recommender systems, significantly outperforming state-of-the-art methods.

Large multimodal foundation models have been extensively utilized for image recognition tasks guided by instructions, yet there remains a scarcity of domain expertise in industrial vibration signal analysis. This paper presents a pipeline named VSLLaVA that leverages a large language model to integrate expert knowledge for identification of signal parameters and diagnosis of faults. Within this pipeline, we first introduce an expert rule-assisted signal generator. The generator merges signal provided by vibration analysis experts with domain-specific parameter identification and fault diagnosis question-answer pairs to build signal-question-answer triplets. Then we use these triplets to apply low-rank adaptation methods for fine-tuning the linear layers of the Contrastive Language-Image Pretraining (CLIP) and large language model, injecting multimodal signal processing knowledge. Finally, the fine-tuned model is assessed through the combined efforts of large language model and expert rules to evaluate answer accuracy and relevance, which showcases enhanced performance in identifying, analyzing various signal parameters, and diagnosing faults. These enhancements indicate the potential of this pipeline to build a foundational model for future industrial signal analysis and monitoring.