English
Related papers

Related papers: Why Does New Knowledge Create Messy Ripple Effects…

200 papers

Recently, using large language models (LLMs) for data augmentation has led to considerable improvements in unsupervised sentence embedding models. However, existing methods encounter two primary challenges: limited data diversity and high…

Computation and Language · Computer Science 2025-10-07 Peichao Lai , Zhengfeng Zhang , Wentao Zhang , Fangcheng Fu , Bin Cui

Knowledge editing aims at updating knowledge of large language models (LLMs) to prevent them from becoming outdated. Existing work edits LLMs at the level of factual knowledge triplets. However, natural knowledge updates in the real world…

Computation and Language · Computer Science 2024-04-23 Hao Peng , Xiaozhi Wang , Chunyang Li , Kaisheng Zeng , Jiangshan Duo , Yixin Cao , Lei Hou , Juanzi Li

Large language models (LLMs) are increasingly used to generate feedback, yet their impact on learning remains underexplored, especially compared to existing feedback methods. This study investigates how on-demand LLM-generated explanatory…

Computation and Language · Computer Science 2025-06-23 Danielle R. Thomas , Conrad Borchers , Shambhavi Bhushan , Erin Gatz , Shivang Gupta , Kenneth R. Koedinger

Agent-compiled knowledge bases provide persistent external knowledge for large language model (LLM) agents in open-ended, knowledge-intensive downstream tasks. Yet their quality is systematically limited by \emph{incompleteness},…

Computation and Language · Computer Science 2026-05-12 Haoyu Huang , Jiaxin Bai , Shujie Liu , Yang Wei , Hong Ting Tsang , Yisen Gao , Zhongwei Xie , Yufei Li , Yangqiu Song

Pre-trained language models (PLMs) like BERT have made significant progress in various downstream NLP tasks. However, by asking models to do cloze-style tests, recent work finds that PLMs are short in acquiring knowledge from unstructured…

Computation and Language · Computer Science 2023-10-12 Cunxiang Wang , Fuli Luo , Yanyang Li , Runxin Xu , Fei Huang , Yue Zhang

Retrieval-augmented generation (RAG) mitigates many problems of fully parametric language models, such as temporal degradation, hallucinations, and lack of grounding. In RAG, the model's knowledge can be updated from documents provided in…

Machine Learning · Computer Science 2024-10-10 Evgenii Kortukov , Alexander Rubinstein , Elisa Nguyen , Seong Joon Oh

Knowledge distillation (KD) is widely used for compressing a teacher model to a smaller student model, reducing its inference cost and memory footprint while preserving model capabilities. However, current KD methods for auto-regressive…

Computation and Language · Computer Science 2024-07-04 Jongwoo Ko , Sungnyun Kim , Tianyi Chen , Se-Young Yun

Traditional knowledge graph completion (KGC) methods rely solely on structural information and struggle with sparsity, while Large Language Models (LLMs) address these limitations through rich world knowledge and strong context modeling.…

Computation and Language · Computer Science 2026-01-30 Bo Xue , Yi Xu , Bolei Ma , Yunchong Song , Jiaxin Ding , Luoyi Fu , Xinbing Wang

Large Language Models (LLMs) have shown remarkable capabilities as AI agents. However, existing methods for enhancing LLM-agent abilities often lack a focus on data quality, leading to inefficiencies and suboptimal results in both…

Machine Learning · Computer Science 2025-02-19 Yunxiao Zhang , Guanming Xiong , Haochen Li , Wen Zhao

In this paper, we focus on the challenging task of reliably estimating factual knowledge that is embedded inside large language models (LLMs). To avoid reliability concerns with prior approaches, we propose to eliminate prompt engineering…

What happens if we train a new Large Language Model (LLM) using data that are at least partially generated by other LLMs? The explosive success of LLMs means that a substantial amount of content online will be generated by LLMs rather than…

Computation and Language · Computer Science 2024-07-26 Jinghui Zhang , Dandan Qiao , Mochen Yang , Qiang Wei

Training large language models (LLMs) typically involves pre-training on massive corpora, only to restart the process entirely when new data becomes available. A more efficient and resource-conserving approach would be continual…

During the pretraining phase, large language models (LLMs) acquire vast amounts of knowledge from extensive text corpora. Nevertheless, in later stages such as fine-tuning and inference, the model may encounter knowledge not covered in the…

Computation and Language · Computer Science 2024-10-10 Bozhou Li , Hao Liang , Yang Li , Fangcheng Fu , Hongzhi Yin , Conghui He , Wentao Zhang

Reward Models (RMs) are crucial for online alignment of language models (LMs) with human preferences. However, RM-based preference-tuning is vulnerable to reward hacking, whereby LM policies learn undesirable behaviors from flawed RMs. By…

Computation and Language · Computer Science 2026-03-05 Daniel Fein , Max Lamparth , Violet Xiang , Mykel J. Kochenderfer , Nick Haber

Large Language Models (LLMs) are increasingly deployed for open-domain question answering, yet their alignment with human perspectives on temporally recent information remains underexplored. We introduce RECOM (Reddit Evaluation for…

Computation and Language · Computer Science 2026-02-13 Pushwitha Krishnappa , Amit Das , Vinija Jain , Tathagata Mukherjee , Aman Chadha

Large Language Models (LLMs) pose a significant risk of safety misalignment after finetuning, as models can be compromised by both explicitly and implicitly harmful data. Even some seemingly benign data can inadvertently steer a model…

Computation and Language · Computer Science 2026-05-15 Zhanhao Hu , Xiao Huang , Patrick Mendoza , Emad A. Alghamdi , Basel Alomair , Raluca Ada Popa , David Wagner

Reinforcement learning (RL) has become the dominant paradigm for improving the performance of language models on complex reasoning tasks. Despite the substantial empirical gains demonstrated by RL-based training methods like GRPO, a…

Artificial Intelligence · Computer Science 2025-10-27 Jiayu Wang , Yifei Ming , Zixuan Ke , Caiming Xiong , Shafiq Joty , Aws Albarghouthi , Frederic Sala

Recent papers show LLMs achieve near-random accuracy in causal relation classification, raising questions about whether such failures arise from limited pretraining exposure or deeper representational gaps. We investigate this under…

Computation and Language · Computer Science 2025-09-25 Oscar Lithgow-Serrano , Vani Kanjirangat , Alessandro Antonucci

Nowadays, the research on Large Vision-Language Models (LVLMs) has been significantly promoted thanks to the success of Large Language Models (LLM). Nevertheless, these Vision-Language Models (VLMs) are suffering from the drawback of…

Computer Vision and Pattern Recognition · Computer Science 2023-11-27 Hongyu Hu , Jiyuan Zhang , Minyi Zhao , Zhenbang Sun

Research on reasoning in language models (LMs) predominantly focuses on improving the correctness of their outputs. But some important applications require modeling reasoning patterns that are incorrect. For example, automated systems that…

Machine Learning · Computer Science 2025-10-14 Alexis Ross , Jacob Andreas