Zobrazeno 1 - 10
of 18
pro vyhledávání: '"Nie, Ercong"'
Large language models (LLMs) possess extensive parametric knowledge, but this knowledge is difficult to update with new information because retraining is very expensive and infeasible for closed-source models. Knowledge editing (KE) has emerged as a
Externí odkaz:
http://arxiv.org/abs/2406.17764
Autor:
Liu, Yongkang, Nie, Ercong, Feng, Shi, Hua, Zheng, Ding, Zifeng, Wang, Daling, Zhang, Yifei, Schütze, Hinrich
Publikováno v:
2024 European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases
Current state-of-the-art dialogue systems heavily rely on extensive training datasets. However, challenges arise in domains where domain-specific training datasets are insufficient or entirely absent. To tackle this challenge, we propose a novel data
Externí odkaz:
http://arxiv.org/abs/2406.09881
Inspired by cognitive neuroscience studies, we introduce a novel `decoding probing' method that uses minimal pairs benchmark (BLiMP) to probe internal linguistic characteristics in neural language models layer by layer. By treating the language model
Externí odkaz:
http://arxiv.org/abs/2403.17299
Autor:
Nie, Ercong, Yuan, Shuzhou, Ma, Bolei, Schmid, Helmut, Färber, Michael, Kreuter, Frauke, Schütze, Hinrich
Despite the predominance of English in their training data, English-centric Large Language Models (LLMs) like GPT-3 and LLaMA display a remarkable ability to perform multilingual tasks, raising questions about the depth and nature of their cross-ling
Externí odkaz:
http://arxiv.org/abs/2402.18397
Large Language Models (LLMs) exhibit strong In-Context Learning (ICL) capabilities when prompts with demonstrations are used. However, fine-tuning still remains crucial to further enhance their adaptability. Prompt-based fine-tuning proves to be an e
Externí odkaz:
http://arxiv.org/abs/2402.11709
Large Language Models (LLMs) possess outstanding capabilities in addressing various natural language processing (NLP) tasks. However, the sheer size of these models poses challenges in terms of storage, training and inference due to the inclusion of
Externí odkaz:
http://arxiv.org/abs/2402.11700
Autor:
Ma, Bolei, Nie, Ercong, Yuan, Shuzhou, Schmid, Helmut, Färber, Michael, Kreuter, Frauke, Schütze, Hinrich
Prompt-based methods have been successfully applied to multilingual pretrained language models for zero-shot cross-lingual understanding. However, most previous studies primarily focused on sentence-level classification tasks, and only a few consider
Externí odkaz:
http://arxiv.org/abs/2401.16589
The remarkable ability of Large Language Models (LLMs) to understand and follow instructions has sometimes been limited by their in-context learning (ICL) performance in low-resource languages. To address this, we introduce a novel approach that leve
Externí odkaz:
http://arxiv.org/abs/2311.06595
The promise of Large Language Models (LLMs) in Natural Language Processing has often been overshadowed by their limited performance in low-resource languages such as Bangla. To address this, our paper presents a pioneering approach that utilizes cros
Externí odkaz:
http://arxiv.org/abs/2311.00587
Pretrained multilingual encoder models can directly perform zero-shot multilingual tasks or linguistic probing by reformulating the input examples into cloze-style prompts. This is accomplished by predicting the probabilities of the label words at th
Externí odkaz:
http://arxiv.org/abs/2310.05069