Zobrazeno 1 - 10
of 2 447
pro vyhledávání: '"HAYES, PETER"'
Retrieval Augmented Generation (RAG) has emerged as an effective solution for mitigating hallucinations in Large Language Models (LLMs). The retrieval stage in RAG typically involves a pre-trained embedding model, which converts queries and passages
Externí odkaz:
http://arxiv.org/abs/2402.12177
As large language models (LLMs) become more capable, fine-tuning techniques for aligning with human intent are increasingly important. A key consideration for aligning these models is how to most effectively use human resources, or model resources in
Externí odkaz:
http://arxiv.org/abs/2402.08114
Autor:
Hayes, Peter John
Dynamic polymer materials have attracted a wide range of interest in different application areas such as the aerospace industry. One of the fundamental challenges in studying crosslinked polymer materials is to understand how the architectural featur
Publikováno v:
High Temperature Materials and Processes, Vol 30, Iss 4, Pp 505-510 (2011)
Externí odkaz:
https://doaj.org/article/88f0c49a8f7e4ae594e521ae5592eff7
Autor:
Zhang, Mingtian, Key, Oscar, Hayes, Peter, Barber, David, Paige, Brooks, Briol, François-Xavier
Score-based divergences have been widely used in machine learning and statistics applications. Despite their empirical success, a blindness problem has been observed when using these for multi-modal distributions. In this work, we discuss the blindne
Externí odkaz:
http://arxiv.org/abs/2209.07396