EMNLP Findings2022
Calibrating Factual Knowledge in Pretrained Language Models
Qingxiu Dong, Damai Dai, Yifan Song, Jingjing Xu, Zhifang Sui, Lei Li
Abstract
Previous literature has proved that Pretrained Language Models (PLMs) can store factual knowledge. However, we find that facts stored in the PLMs are not always correct. It motivates us to explore a fundamental question: How do we calibrate factual knowledge in PLMs without re-training from scratch? In this work, we propose a simple and lightweight method CA-LINET to achieve this goal. To be specific, we first detect whether PLMs can learn the right facts via a contrastive score between right and fake facts. If not, we then use a lightweight method to add and adapt new parameters to specific factual texts. Experiments on the knowledge probing task show the calibration effectiveness and efficiency. In addition, through closed-book question answering, we find that the calibrated PLM possesses knowledge generalization ability after fine-tuning. Beyond the calibration performance, we further investigate and visualize the knowledge calibration mechanism. The code and data are available at https://github.com/dqxiu/CaliNet.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
64 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- Can We Edit Factual Knowledge by In-Context Learning? cites Fast Model Editing at Scale
- Can We Edit Factual Knowledge by In-Context Learning? cites Language Models are Few-Shot Learners
- MELO: Enhancing Model Editing with Neuron-Indexed Dynamic LoRA cites Editing Large Language Models: Problems, Methods, and Opportunities
- IRCAN: Mitigating Knowledge Conflicts in LLM Generation via Identifying and Reweighting Context-Aware Neurons cites Fast Model Editing at Scale
- Fast Model Editing at Scale cites Language Models are Few-Shot Learners
- Editing Large Language Models: Problems, Methods, and Opportunities cites Can We Edit Factual Knowledge by In-Context Learning?
- Editing Large Language Models: Problems, Methods, and Opportunities cites Language Models are Few-Shot Learners
Cited by top-tier papers64
- Is Stack Overflow Obsolete? An Empirical Study of the Characteristics of ChatGPT Answers to Stack Overflow QuestionsSamia Kabir, David N. Udo-Imeh, Bonan Kou, Tianyi ZhangCHI 2024 · 149 citations
- MELO: Enhancing Model Editing with Neuron-Indexed Dynamic LoRALang Yu, Qin Chen, Jie Zhou, Liang HeAAAI 2024 · 96 citations
- Editing Large Language Models: Problems, Methods, and OpportunitiesYunzhi Yao, Peng Wang, Bozhong Tian, Siyuan Cheng et al.EMNLP 2023 · 83 citations
- IRCAN: Mitigating Knowledge Conflicts in LLM Generation via Identifying and Reweighting Context-Aware NeuronsDan Shi, Renren Jin, Tianhao Shen, Weilong Dong et al.NeurIPS 2024 · 44 citations
- Can We Edit Factual Knowledge by In-Context Learning?Ce Zheng, Lei Li, Qingxiu Dong, Yuxuan Fan et al.EMNLP 2023 · 40 citations
Builds on10
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah et al.NeurIPS 2020 · 64,255 citations
- Fast Model Editing at ScaleEric Mitchell, Charles Lin, Antoine Bosselut, Chelsea Finn et al.ICLR 2022 · 527 citations
- K-Adapter: Infusing Knowledge into Pre-Trained Models with AdaptersRuize Wang, Duyu Tang, Nan Duan, Zhongyu Wei et al.ACL Findings 2021 · 473 citations
- Inducing Relational Knowledge from BERTZied Bouraoui, José Camacho-Collados, Steven SchockaertAAAI 2020 · 183 citations
- Don't Stop Pretraining: Adapt Language Models to Domains and TasksSuchin Gururangan, Ana Marasovic, Swabha Swayamdipta, Kyle Lo et al.ACL 2020 · 93 citations
Related papers
- Give Me the Facts! A Survey on Factual Knowledge Probing in Pre-trained Language ModelsPaul Youssef, Osman Alperen Koras, Meijie Li, Jörg Schlötterer et al.EMNLP Findings 2023 · 9 citations
- Cross-Lingual Consistency of Factual Knowledge in Multilingual Language ModelsJirui Qi, Raquel Fernández, Arianna BisazzaEMNLP 2023 · 9 citations
- Pre-training Language Models with Deterministic Factual KnowledgeShaobo Li, Xiaoguang Li, Lifeng Shang, Chengjie Sun et al.EMNLP 2022 · 13 citations
- Knowledge-Augmented Language Model VerificationJinheon Baek, Soyeong Jeong, Minki Kang, Jong C. Park et al.EMNLP 2023 · 12 citations
- Understanding Finetuning for Factual Knowledge ExtractionGaurav Rohit Ghosal, Tatsunori Hashimoto, Aditi RaghunathanICML 2024 · 36 citations
