English
Related papers

Related papers: IF-GUIDE: Influence Function-Guided Detoxification…

200 papers

Recommendation unlearning is an emerging task to serve users for erasing unusable data (e.g., some historical behaviors) from a well-trained recommender model. Existing methods process unlearning requests by fully or partially retraining…

Information Retrieval · Computer Science 2024-11-01 Yang Zhang , Zhiyu Hu , Yimeng Bai , Jiancan Wu , Qifan Wang , Fuli Feng

How can we explain the influence of training data on black-box models? Influence functions (IFs) offer a post-hoc solution by utilizing gradients and Hessians. However, computing the Hessian for an entire dataset is resource-intensive,…

Machine Learning · Computer Science 2025-11-03 Jungyeon Koh , Hyeonsu Lyu , Jonggyu Jang , Hyun Jong Yang

Despite significant progress, recent studies indicate that current large language models (LLMs) may still capture dataset biases and utilize them during inference, leading to the poor generalizability of LLMs. However, due to the diversity…

Computation and Language · Computer Science 2025-05-28 Zhouhao Sun , Xiao Ding , Li Du , Yunpeng Xu , Yixuan Ma , Yang Zhao , Bing Qin , Ting Liu

Large language models (LM) generate remarkably fluent text and can be efficiently adapted across NLP tasks. Measuring and guaranteeing the quality of generated text in terms of safety is imperative for deploying LMs in the real world; to…

Large Language Models (LLMs) have achieved remarkable success, where instruction tuning is the critical step in aligning LLMs with user intentions. In this work, we investigate how the instruction tuning adjusts pre-trained models with a…

Computation and Language · Computer Science 2024-04-05 Xuansheng Wu , Wenlin Yao , Jianshu Chen , Xiaoman Pan , Xiaoyang Wang , Ninghao Liu , Dong Yu

Pretrained neural language models (LMs) are prone to generating racist, sexist, or otherwise toxic language which hinders their safe deployment. We investigate the extent to which pretrained LMs can be prompted to generate toxic language,…

Computation and Language · Computer Science 2020-09-29 Samuel Gehman , Suchin Gururangan , Maarten Sap , Yejin Choi , Noah A. Smith

Curriculum learning, a training technique where data is presented to the model in order of example difficulty (e.g., from simpler to more complex documents), has shown limited success for pre-training language models. In this work, we…

Computation and Language · Computer Science 2025-09-29 Loris Schoenegger , Lukas Thoma , Terra Blevins , Benjamin Roth

Influence functions (IF) have been seen as a technique for explaining model predictions through the lens of the training data. Their utility is assumed to be in identifying training examples "responsible" for a prediction so that, for…

Machine Learning · Computer Science 2023-05-29 Andrea Schioppa , Katja Filippova , Ivan Titov , Polina Zablotskaia

Federated Instruction Tuning (FIT) enables collaborative instruction tuning of large language models across multiple organizations (clients) in a cross-silo setting without requiring the sharing of private instructions. Recent findings on…

Cryptography and Security · Computer Science 2026-03-03 Haodong Zhao , Jinming Hu , Zhaomin Wu , Zongru Wu , Wei Du , Junyi Hou , Caibei Zhao , Zhuosheng Zhang , Bingsheng He , Gongshen Liu

Language models (LMs) often exhibit undesirable text generation behaviors, including generating false, toxic, or irrelevant outputs. Reinforcement learning from human feedback (RLHF) - where human preference judgments on LM outputs are…

Computation and Language · Computer Science 2023-10-31 Zeqiu Wu , Yushi Hu , Weijia Shi , Nouha Dziri , Alane Suhr , Prithviraj Ammanabrolu , Noah A. Smith , Mari Ostendorf , Hannaneh Hajishirzi

Selecting appropriate training data is crucial for effective instruction fine-tuning of large language models (LLMs), which aims to (1) elicit strong capabilities, and (2) achieve balanced performance across a diverse range of tasks.…

Computation and Language · Computer Science 2025-01-22 Qirun Dai , Dylan Zhang , Jiaqi W. Ma , Hao Peng

Large-scale black-box models have become ubiquitous across numerous applications. Understanding the influence of individual training data sources on predictions made by these models is crucial for improving their trustworthiness. Current…

Machine Learning · Computer Science 2024-06-21 Myeongseob Ko , Feiyang Kang , Weiyan Shi , Ming Jin , Zhou Yu , Ruoxi Jia

When trying to gain better visibility into a machine learning model in order to understand and mitigate the associated risks, a potentially valuable source of evidence is: which training examples most contribute to a given behavior?…

Large Language Models (LLMs) can acquire unintended biases from seemingly benign training data even without explicit cues or malicious content. Existing methods struggle to detect such risks before fine-tuning, making post hoc evaluation…

Machine Learning · Computer Science 2026-02-05 Mengru Wang , Zhenqian Xu , Junfeng Fang , Yunzhi Yao , Shumin Deng , Huajun Chen , Ningyu Zhang

Stealthy data poisoning during fine-tuning can backdoor large language models (LLMs), threatening downstream safety. Existing detectors either use classifier-style probability signals--ill-suited to generation--or rely on rewriting, which…

Computation and Language · Computer Science 2025-11-13 Jinwen Chen , Hainan Zhang , Fei Sun , Qinnan Zhang , Sijia Wen , Ziwei Wang , Zhiming Zheng

Data filtering strategies are a crucial component to develop safe Large Language Models (LLM), since they support the removal of harmful contents from pretraining datasets. There is a lack of research on the actual impact of these…

Computation and Language · Computer Science 2026-03-24 Marco Antonio Stranisci , Christian Hardmeier

Training data attribution (TDA) methods aim to attribute model outputs back to specific training examples, and the application of these methods to large language model (LLM) outputs could significantly advance model transparency and data…

Computation and Language · Computer Science 2024-12-24 Tyler A. Chang , Dheeraj Rajagopal , Tolga Bolukbasi , Lucas Dixon , Ian Tenney

Instruction tuning plays a crucial role in shaping the outputs of language models (LMs) to desired styles. In this work, we propose a simple yet effective method, Instruction Modelling (IM), which trains LMs by applying a loss function to…

Computation and Language · Computer Science 2024-10-04 Zhengyan Shi , Adam X. Yang , Bin Wu , Laurence Aitchison , Emine Yilmaz , Aldo Lipani

Large language models (LLMs) have become integral to various real-world applications, leveraging massive, web-sourced datasets like Common Crawl, C4, and FineWeb for pretraining. While these datasets provide linguistic data essential for…

Computation and Language · Computer Science 2025-08-14 Sai Krishna Mendu , Harish Yenala , Aditi Gulati , Shanu Kumar , Parag Agrawal

Language models trained on large-scale unfiltered datasets curated from the open web acquire systemic biases, prejudices, and harmful views from their training data. We present a methodology for programmatically identifying and removing…

Computation and Language · Computer Science 2021-11-30 Helen Ngo , Cooper Raterink , João G. M. Araújo , Ivan Zhang , Carol Chen , Adrien Morisot , Nicholas Frosst