9 papers
From Early Encoding to Late Suppression: Interpreting LLMs on Character Counting Tasks
Ayan Datta, Mounika Marreddy, Alexander Mehler +2
Large language models (LLMs) exhibit failures on elementary symbolic tasks such as character counting in a word, despite excelling on complex benchmarks. Although this limitation h…
Explanation Generation for Contradiction Reconciliation with LLMs
Jason Chan, Zhixue Zhao, Robert Gaizauskas
Existing NLP work commonly treats contradictions as errors to be resolved by choosing which statements to accept or discard. Yet a key aspect of human reasoning in social and profe…
Can Fine-Tuning Erase Your Edits? On the Fragile Coexistence of Knowledge Editing and Adaptation
Yinjie Cheng, Paul Youssef, Christin Seifert +2
Knowledge editing (KE) offers a lightweight alternative to retraining for updating large language models (LLMs). Meanwhile, fine-tuning remains the default operation for adapting L…
Tracing and Reversing Edits in LLMs
Paul Youssef, Zhixue Zhao, Christin Seifert +1
Knowledge editing methods (KEs) are a cost-effective way to update the factual content of large language models (LLMs), but they pose a dual-use risk. While KEs are beneficial for…
Position: Editing Large Language Models Poses Serious Safety Risks
Paul Youssef, Zhixue Zhao, Daniel Braun +2
Large Language Models (LLMs) contain large amounts of facts about the world. These facts can become outdated over time, which has led to the development of knowledge editing method…
How to Make LLMs Forget: On Reversing In-Context Knowledge Edits
Paul Youssef, Zhixue Zhao, Jörg Schlötterer +1
In-context knowledge editing (IKE) enables efficient modification of large language model (LLM) outputs without parameter changes and at zero-cost. However, it can be misused to ma…
On the Limitations of Language Targeted Pruning: Investigating the Calibration Language Impact in Multilingual LLM Pruning
Simon Kurz, Jian-Jia Chen, Lucie Flek +1
Recent advances in large language model (LLM) pruning have shown state-of-the-art (SotA) compression results in post-training and retraining-free settings while maintaining high pr…
Has this Fact been Edited? Detecting Knowledge Edits in Language Models
Paul Youssef, Zhixue Zhao, Christin Seifert +1
Knowledge editing methods (KEs) can update language models' obsolete or inaccurate knowledge learned from pre-training. However, KEs can be used for malicious applications, e.g., i…
ReAGent: A Model-agnostic Feature Attribution Method for Generative Language Models
Zhixue Zhao, Boxuan Shan
Feature attribution methods (FAs), such as gradients and attention, are widely employed approaches to derive the importance of all input features to the model predictions. Existing…