按年被提到
20242024 · 1 次2026
大模型与语言模型›大模型预训练与缩放·arXiv:2405.05417 · cs.CL

Fishing for Magikarp: Automatically Detecting Under-trained Tokens in Large Language Models

Sander Land, Max Bartolo · 2024
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2024-05-12 · 最近 2025-06-25
2不同的人
2活跃月份
2讨论串
12被引(OpenAlex)
arXiv 摘要(候选层不写导读)

The disconnect between tokenizer creation and model training in language models allows for specific inputs, such as the infamous SolidGoldMagikarp token, to induce unwanted model behaviour. Although such `glitch tokens', tokens present in the tokenizer vocabulary but that are nearly or entirely absent during model training, have been observed across various models, a reliable method to identify and address them has been missing. We present a comprehensive analysis of Large Language Model tokenizers, specifically targeting this issue of detecting under-trained tokens. Through a combination of tokenizer analysis, model weight-based indicators, and prompting techniques, we develop novel and effective methods for automatically detecting these problematic tokens. Our findings demonstrate the prevalence of such tokens across a diverse set of models and provide insights into improving the efficiency and safety of language models.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
评论cschmidt · 2025-06

“Virtually all current tokenization schemes do work at the raw byte level, not the utf-8 character. They do this to avoid the Out of Vocabulary (OOV) or unknown token problem. In older models, if you came across something in the data you can't tokenize, you add a <UNK>. But tokenization should be exactly reversible, so now people use subword tokenizers inc…”

The bitter lesson is coming for tokenization原评论 ↗
被提交为 story · 1 次