paper-with-me

홈 › Papers

Token Alignment via Character Matching for Subword Completion

2024-03-13 · Ben Athiwaratkun, Shiqi Wang, Mingyue Shang, Yuchen Tian, Zijian Wang, Sujan Kumar Gonugondla, Sanjay Krishna Gouda, Rob Kwiatowski, Ramesh Nallapati, Bing Xiang

Generative models, widely utilized in various applications, can often struggle with prompts corresponding to partial tokens. This struggle stems from tokenization, where partial tokens fall out of distribution during inference, leading to incorrect or nonsensical outputs. This paper examines a technique to alleviate the tokenization artifact on text completion in generative models, maintaining performance even in regular non-subword cases. The method, termed token alignment, involves backtracking to the last complete tokens and ensuring the model's generation aligns with the prompt. This approach showcases marked improvement across many partial token scenarios, including nuanced cases like space-prefix and partial indentation, with only a minor time increase. The technique and analysis detailed in this paper contribute to the continuous advancement of generative models in handling partial inputs, bearing relevance for applications like code completion and text autocompletion.

📄 PDF Abstract BibTeX arXiv:2403.08688

Code (0)

등록된 구현이 없습니다.

Tasks

Code Completion

Similar Papers 제목 키워드 기반

StreamAlign: Streaming Text-Aligned Speech Tokenization

2026-09-09 · Kang-wook Kim, Jinyoung Park, Jinsoo Kim, Sehun Lee 외 arxiv

Text-aligned speech tokenization methods have emerged to better align speech tokens with LLM token spaces, enabling more effective utilization of pretrained LLMs. However, they rely on offline automatic speech recognitio…

Speech Recognition

Subword Language Model for Query Auto-Completion

2019-09-02 · IJCNLP 2019 11 · Gyuwan Kim

Current neural query auto-completion (QAC) systems rely on character-level language models, but they slow down when queries are long. We present how to utilize subword language models for the fast and accurate generation…

Language ModelingLanguage ModellingmodelReranking

MaxMatch-Dropout: Subword Regularization for WordPiece

2022-09-09 · COLING 2022 10 · Tatsuya Hiraoka

We present a subword regularization method for WordPiece, which uses a maximum matching algorithm for tokenization. The proposed method, MaxMatch-Dropout, randomly drops words in a search using the maximum matching algor…

Machine TranslationText ClassificationTranslation

Charformer: Fast Character Transformers via Gradient-based Subword Tokenization

2021-06-23 · ICLR 2022 4 · Yi Tay, Vinh Q. Tran, Sebastian Ruder, Jai Gupta 외

State-of-the-art models in natural language processing rely on separate rigid subword tokenization algorithms, which limit their generalization ability and adaptation to new settings. In this paper, we propose a new mode…

Inductive BiasLinguistic AcceptabilityNatural Language InferenceParaphrase Identification+2

Char2Subword: Extending the Subword Embedding Space Using Robust Character Compositionality

2020-10-24 · Findings (EMNLP) 2021 11 · Gustavo Aguilar, Bryan McCann, Tong Niu, Nazneen Rajani 외

Byte-pair encoding (BPE) is a ubiquitous algorithm in the subword tokenization process of language models as it provides multiple benefits. However, this process is solely based on pre-training data statistics, making it…