Factual Error Detection (FED), which is the task of identifying factually incorrect spans in a given text, has long been recognized as an important research problem. However, with the rapid rise of large language models (LLMs), research attention has shifted toward factual errors specific to LLM-generated text (hallucinations) and their detection. As a result, the detection of factual errors in human-written text has been relatively neglected. To address this gap, we first distill a taxonomy of human-induced factual errors by analyzing corrections of newspaper articles, a representative source of text that is guaranteed to be human-written and contains few grammatical errors. Our analysis revealed that there are characteristic categories such as kanji misconversions and numeral classifier errors, which are not focused in existing hallucination benchmarks. Based on the taxonomy, we then evaluate the FED capability of vanilla LLMs on synthesized realistic test cases and real corrections. Experimental results demonstrated that even high-performance LLMs such as GPT-5.4 achieved only a word-level F1 score of 52% on the synthetic evaluation data, highlighting the task difficulty. Furthermore, a detailed analysis by detection difficulty revealed the current state of FED.

This study presents a difficulty-annotated corpus of Japanese newspaper articles. The annotation in our corpus has two key features. First, it is hierarchical, with difficulty levels labeled at the word, sentence, and document levels, respectively. Second, it is subjective, with 99 readers from diverse backgrounds labeling the difficulty levels based on their own perceptions. Detailed analysis of our corpus revealed correlations in difficulty levels across linguistic units, such as between words and sentences, and sentences and documents. We also found that text difficulty varies depending on reader attributes such as age and topics of interest. Furthermore, we trained Transformer encoders on this corpus and developed a personalized difficulty estimation model that takes into account both linguistic and attribute information. Experimental results demonstrate that incorporating reader attributes improves the performance of difficulty estimation and that multi-task learning across language units is beneficial. Our difficulty annotations are publicly available.

News articles often contain omitted temporal expressions, such as day-only or month-only mentions, which must be interpreted with reference to the publication date. When such articles are indexed or processed as standalone text in search and retrieval-augmented generation (RAG) systems, these omissions can cause temporal mismatches and unstable interpretation by large language models. We focus on reproducing omitted temporal expressions as concrete dates or intervals using the publication date as external context before the articles are indexed for search and RAG applications. Specifically, building on established temporal-expression extraction and normalization techniques and informed by a manual analysis of Japanese news articles, we propose jaROTE, a rule-based pipeline for Japanese news. Experiments on two news corpora demonstrate that jaROTE achieves high performance, and remains competitive with LLMs while providing a fast, low-cost pipeline. We further show that temporal reproduction improves time-constrained lexical retrieval, demonstrating the practical value of publication-date-grounded normalization for Japanese news retrieval.

We propose Fast-MIA (https://github.com/Nikkei/fast-mia), a Python library for efficiently evaluating membership inference attacks (MIA) against large language models (LLMs). MIA has emerged as a crucial technique for auditing privacy risks and copyright infringement in LLMs. However, computational demands have grown substantially: recent methods rely on repeated inference, while practical auditing requires large-scale evaluation. Progress is further hindered by existing implementations that execute methods independently, redundantly computing shared intermediate results such as log-probabilities. To address these challenges, Fast-MIA combines two strategies: (1) high-throughput batch inference via vLLM, achieving approximately 5× speedup, and (2) a cross-method caching architecture that computes intermediate results once and shares them across methods.The library includes representative MIA methods under a unified framework, integrates with established benchmarks, and supports flexible YAML configuration.We release Fast-MIA under the Apache License 2.0 to support scalable and reproducible MIA research.

Despite the growing concern about memorization of training data using large language models (LLMs), there has been insufficient analysis under conditions using non-English or industry-specific corpora.This study focuses on continual pre-training, a common approach in building non-English LLMs, and quantifies memorization of training data.Specifically, we trained two models based on Llama 3 using Japanese Wikipedia (general) and Japanese financial news articles (industry-specific).Experiments showed a tendency for the amount of memorization to increase as training progressed, similar to the empirical findings for English.This trend was clear in the industry-specific corpus, suggesting potential risks when using valuable, non-general industry corpora.We also identified issues specific to Japanese, and emphasized the importance of analysis other than in English.

Dominant pre-trained language models (PLMs) have demonstrated the potential risk of memorizing and outputting the training data. While this concern has been discussed mainly in English, it is also practically important to focus on domain-specific PLMs. In this study, we pre-trained domain-specific GPT-2 models using a limited corpus of Japanese newspaper articles and evaluated their behavior. Experiments replicated the empirical finding that memorization of PLMs is related to the duplication in the training data, model size, and prompt length, in Japanese the same as in previous English studies. Furthermore, we attempted membership inference attacks, demonstrating that the training data can be detected even in Japanese, which is the same trend as in English. The study warns that domain-specific PLMs, sometimes trained with valuable private data, can ”copy and paste” on a large scale.

As the deployment of pre-trained language models (PLMs) expands, pressing security concerns have arisen regarding the potential for malicious extraction of training data, posing a threat to data privacy. This study is the first to provide a comprehensive survey of training data extraction from PLMs.Our review covers more than 100 key papers in fields such as natural language processing and security. First, preliminary knowledge is recapped and a taxonomy of various definitions of memorization is presented. The approaches for attack and defense are then systemized. Furthermore, the empirical findings of several quantitative studies are highlighted. Finally, future research directions based on this review are suggested.

This paper describes our system in SemEval-2022 Task 8, where participants were required to predict the similarity of two multilingual news articles. In the task of pairwise sentence and document scoring, there are two main approaches: Cross-Encoder, which inputs pairs of texts into a single encoder, and Bi-Encoder, which encodes each input independently. The former method often achieves higher performance, but the latter gave us a better result in SemEval-2022 Task 8. This paper presents our exploration of BERT-based Bi-Encoder approach for this task, and there are several findings such as pretrained models, pooling methods, translation, data separation, and the number of tokens. The weighted average ensemble of the four models achieved the competitive result and ranked in the top 12.

Word embeddings and pre-trained language models have become essential technical elements in natural language processing. While the general practice is to use or fine-tune publicly available models, there are significant advantages in creating or pre-training unique models that match the domain. The performance of the models degrades as language changes or evolves continuously, but the high cost of model building inhibits regular re-training, especially for the language models. This study proposes an efficient way to detect time-series performance degradation of word embeddings and pre-trained language models by calculating the degree of semantic shift. Monitoring performance through the proposed method supports decision-making as to whether a model should be re-trained. The experiments demonstrated that the proposed method can identify time-series performance degradation in two datasets, Japanese and English. The source code is available at https://github.com/Nikkei/semantic-shift-stability.