Related papers: Data Mixture Inference: What do BPE Tokenizers Rev…

Tokenizer adaptation plays an important role in adapting pre-trained language models to new domains or languages. In this work, we address two complementary aspects of this process: vocabulary extension and pruning. The common approach to…

Pretraining data of large language models composes multiple domains (e.g., web texts, academic papers, codes), whose mixture proportions crucially impact the competence of outcome models. While existing endeavors rely on heuristics or…

Natural language processing (NLP) enables the understanding and generation of meaningful human language, typically using a pre-trained complex architecture on a large dataset to learn the language and next fine-tune its weights to implement…

Large language models have demonstrated remarkable capabilities across various tasks, primarily attributed to the utilization of diversely sourced data. However, the impact of pretraining data composition on model performance remains poorly…

MixUp is a data augmentation strategy where additional samples are generated during training by combining random pairs of training samples and their labels. However, selecting random pairs is not potentially an optimal choice. In this work,…

Optimizing data mixtures is essential for unlocking the full potential of large language models (LLMs), yet identifying the optimal composition remains computationally prohibitive due to reliance on heuristic trials or expensive proxy…

We propose a method to optimize language model pre-training data mixtures through efficient approximation of the cross-entropy loss corresponding to each candidate mixture via a Mixture of Data Experts (MDE). We use this approximation as a…

Tokenization is the first step in modern neural language model pipelines where an input text is converted to a sequence of subword tokens. We introduce from first principles a finite-state transduction framework which can efficiently encode…

Recent work has shown that Large Language Models (LLMs) can unintentionally leak sensitive information present in their training data. In this paper, we present Model Perturbations (MoPe), a new method to identify with high confidence if a…

Tokenization is a necessary component within the current architecture of many language mod-els, including the transformer-based large language models (LLMs) of Generative AI, yet its impact on the model's cognition is often overlooked. We…

The best performing transformer-based language models use subword tokenization techniques, such as Byte-Pair-Encoding (BPE). However, these approaches often overlook linguistic principles, such as morphological segmentation, which we…

The pre-trained language models have achieved great successes in various natural language understanding (NLU) tasks due to its capacity to capture the deep contextualized information in text by pre-training on large-scale corpora. One of…

Mixture-of-Experts (MoE) models have shown remarkable capability in instruction tuning, especially when the number of tasks scales. However, previous methods simply merge all training tasks (e.g. creative writing, coding, and mathematics)…

Computing next-token likelihood ratios between two language models (LMs) is a standard task in training paradigms such as knowledge distillation. Since this requires both models to share the same probability space, it becomes challenging…

Gender-inclusive NLP research has documented the harmful limitations of gender binary-centric large language models (LLM), such as the inability to correctly use gender-diverse English neopronouns (e.g., xe, zir, fae). While data scarcity…

Despite it being the cornerstone of BPE, the most common tokenization algorithm, the importance of compression in the tokenization process is still unclear. In this paper, we argue for the theoretical importance of compression, that can be…

Byte Pair Encoding (BPE) tokenizers, widely used in Large Language Models, face challenges in multilingual settings, including penalization of non-Western scripts and the creation of tokens with partial UTF-8 sequences. Pretokenization,…

Tokenization is a crucial step in processing protein sequences for machine learning models, as proteins are complex sequences of amino acids that require meaningful segmentation to capture their functional and structural properties.…

Large language models (LLMs) for code rely on subword tokenizers, such as byte-pair encoding (BPE), learned from mixed natural language text and programming language code but driven by statistics rather than grammar. As a result,…

Traditional greedy tokenization methods have been a critical step in Natural Language Processing (NLP), influencing how text is converted into tokens and directly impacting model performance. While subword tokenizers like Byte-Pair Encoding…