Vocabulary Expansion of Chat Models with Unlabeled Target Language Data

• URL: http://arxiv.org/abs/2412.11704v2
• Date: Wed, 18 Dec 2024 12:29:11 GMT
• Title: Vocabulary Expansion of Chat Models with Unlabeled Target Language Data
• Authors: Atsuki Yamaguchi, Terufumi Morishita, Aline Villavicencio, Nikolaos Aletras,
• Abstract summary: Chat models (i.e. language models trained to follow instructions through conversation with humans) outperform base models (i.e. trained solely on unlabeled data) in both conversation and general task-solving abilities.<n>These models are generally English-centric and require further adaptation for languages that are underrepresented in or absent from their training data.<n>We propose post-hoc techniques that inject information from the source model without requiring any further training. Experiments reveal the effectiveness of our methods, helping the adapted models to achieve performance improvements in 87% of cases.
• Score: 38.341705137026985
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Chat models (i.e. language models trained to follow instructions through conversation with humans) outperform base models (i.e. trained solely on unlabeled data) in both conversation and general task-solving abilities. These models are generally English-centric and require further adaptation for languages that are underrepresented in or absent from their training data. A common technique for adapting base models is to extend the model's vocabulary with target language tokens, i.e. vocabulary expansion (VE), and then continually pre-train it on language-specific data. Using chat data is ideal for chat model adaptation, but often, either this does not exist or is costly to construct. Alternatively, adapting chat models with unlabeled data is a possible solution, but it could result in catastrophic forgetting. In this paper, we investigate the impact of using unlabeled target language data for VE on chat models for the first time. We first show that off-the-shelf VE generally performs well across target language tasks and models in 71% of cases, though it underperforms in scenarios where source chat models are already strong. To further improve adapted models, we propose post-hoc techniques that inject information from the source model without requiring any further training. Experiments reveal the effectiveness of our methods, helping the adapted models to achieve performance improvements in 87% of cases.

Related papers

• Hidden in Plain Sight: Exploring Chat History Tampering in Interactive Language Models [12.920884182101142]
  Large Language Models (LLMs) have become prevalent in real-world applications, exhibiting impressive text generation performance. To behave interactively, LLM-based chat systems must integrate prior chat history as context into their inputs, following a pre-defined structure. This paper introduces a systematic methodology to inject user-supplied history into LLM conversations without any prior knowledge of the target model.
  arXiv  Detail & Related papers  (2024-05-30T16:36:47Z)
• Why Not Transform Chat Large Language Models to Non-English? [57.16587777261422]
  The scarcity of non-English data limits the development of non-English large language models (LLMs) TransLLM divides the transfer problem into some common sub-tasks with the translation chain-of-thought. Our method, using only single-turn data, outperforms strong baselines and ChatGPT on multi-turn benchmark MT-bench.
  arXiv  Detail & Related papers  (2024-05-22T18:53:25Z)
• ChatEL: Entity Linking with Chatbots [11.944348800783834]
  ChatEL is a three-step framework to prompt Large Language Models to return accurate results. Overall the ChatEL framework improves the average F1 performance across 10 datasets by more than 2%.
  arXiv  Detail & Related papers  (2024-02-20T20:52:57Z)
• Efficiently Adapting Pretrained Language Models To New Languages [9.33333013114014]
  Recent large language models (LLM) exhibit sub-optimal performance on low-resource languages. We study how to efficiently adapt any existing pretrained LLM to a new language without running into these issues.
  arXiv  Detail & Related papers  (2023-11-09T20:59:08Z)
• Robustifying Language Models with Test-Time Adaptation [17.96043752001886]
  Large-scale language models achieved state-of-the-art performance over a number of language tasks. They fail on adversarial language examples, which are sentences optimized to fool the language models but with similar semantic meanings for humans. We show that we can reverse many language adversarial attacks by adapting the input sentence with predictions from masked words.
  arXiv  Detail & Related papers  (2023-10-29T22:37:54Z)
• Chat Vector: A Simple Approach to Equip LLMs with Instruction Following and Model Alignment in New Languages [40.37822682459469]
  We introduce the concept of $textitchat vector$ to equip pre-trained language models with instruction following and human value alignment. By simply adding the chat vector to a continual pre-trained model's weights, we can endow the model with chat capabilities without the need for languages.
  arXiv  Detail & Related papers  (2023-10-07T13:34:21Z)
• Qwen Technical Report [132.54304067403922]
  We introduce Qwen, the first installment of our large language model series. Qwen is the base pretrained language models, and Qwen-Chat, the chat models finetuned with human alignment techniques. We have also developed coding-specialized models, Code-Qwen and Code-Qwen-Chat, as well as mathematics-focused models, Math-Qwen-Chat.
  arXiv  Detail & Related papers  (2023-09-28T17:07:49Z)
• ChatGPT Beyond English: Towards a Comprehensive Evaluation of Large Language Models in Multilingual Learning [70.57126720079971]
  Large language models (LLMs) have emerged as the most important breakthroughs in natural language processing (NLP) This paper evaluates ChatGPT on 7 different tasks, covering 37 diverse languages with high, medium, low, and extremely low resources. Compared to the performance of previous models, our extensive experimental results demonstrate a worse performance of ChatGPT for different NLP tasks and languages.
  arXiv  Detail & Related papers  (2023-04-12T05:08:52Z)
• Baize: An Open-Source Chat Model with Parameter-Efficient Tuning on Self-Chat Data [101.63682141248069]
  Chat models, such as ChatGPT, have shown impressive capabilities and have been rapidly adopted across numerous domains. We propose a pipeline that can automatically generate a high-quality multi-turn chat corpus by leveraging ChatGPT. We employ parameter-efficient tuning to enhance LLaMA, an open-source large language model.
  arXiv  Detail & Related papers  (2023-04-03T17:59:09Z)
• Contrastive Alignment of Vision to Language Through Parameter-Efficient Transfer Learning [60.26952378997713]
  Contrastive vision-language models (e.g. CLIP) are created by updating all the parameters of a vision model and language model through contrastive training. We show that a minimal set of parameter updates ($$7%) can achieve the same performance as full-model training. We describe a series of experiments: we show that existing knowledge is conserved more strongly in parameter-efficient training.
  arXiv  Detail & Related papers  (2023-03-21T14:12:08Z)
• Efficient Language Model Training through Cross-Lingual and Progressive Transfer Learning [0.7612676127275795]
  Most Transformer language models are pretrained on English text. As model sizes grow, the performance gap between English and other languages increases even further. We introduce a cross-lingual and progressive transfer learning approach, called CLP-Transfer.
  arXiv  Detail & Related papers  (2023-01-23T18:56:12Z)
• Code Switching Language Model Using Monolingual Training Data [0.0]
  Training a code-switching (CS) language model using only monolingual data is still an ongoing research problem. In this work, an RNN language model is trained using alternate batches from only monolingual English and Spanish data. Results were consistently improved using mean square error (MSE) in the output embeddings of RNN based language model.
  arXiv  Detail & Related papers  (2020-12-23T08:56:39Z)
• Pre-Training a Language Model Without Human Language [74.11825654535895]
  We study how the intrinsic nature of pre-training data contributes to the fine-tuned downstream performance. We find that models pre-trained on unstructured data beat those trained directly from scratch on downstream tasks. To our great astonishment, we uncover that pre-training on certain non-human language data gives GLUE performance close to performance pre-trained on another non-English language.
  arXiv  Detail & Related papers  (2020-12-22T13:38:06Z)
• Detecting and Exorcising Statistical Demons from Language Models with Anti-Models of Negative Data [13.392212395386933]
  We find that within a model family, as the number of parameters, training epochs, and data set size increase, so does a model's ability to generalize to negative n-gram data. We propose a form of inductive bias that attenuates such undesirable signals with negative data distributions automatically learned from positive data.
  arXiv  Detail & Related papers  (2020-10-22T16:45:32Z)
• Comparison of Interactive Knowledge Base Spelling Correction Models for Low-Resource Languages [81.90356787324481]
  Spelling normalization for low resource languages is a challenging task because the patterns are hard to predict. This work shows a comparison of a neural model and character language models with varying amounts on target language data. Our usage scenario is interactive correction with nearly zero amounts of training examples, improving models as more data is collected.
  arXiv  Detail & Related papers  (2020-10-20T17:31:07Z)
• Grounded Compositional Outputs for Adaptive Language Modeling [59.02706635250856]
  A language model's vocabulary$-$typically selected before training and permanently fixed later$-$affects its size. We propose a fully compositional output embedding layer for language models. To our knowledge, the result is the first word-level language model with a size that does not depend on the training vocabulary.
  arXiv  Detail & Related papers  (2020-09-24T07:21:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.