Explore >> Select a destination
|
You are here 
|
deepmind.google | ||
| | | | |
irisvanrooijcogsci.com
|
|
| | | | | Three weeks ago, I wrote a blogpost about how ChatGPT is a "stochastic parrot" (a term coined by Bender, Gebru, McMillan-Major, & Shmitchell, 2021; see also this video for an explanation) and when used for academic (and other) writing constitutes automated plagiarism. My aim was to bring the discussion down to earth and prevent that... | |
| | | | |
www.frontiersin.org
|
|
| | | | | [AI summary] The article discusses the importance of sociolinguistic insights in the development and evaluation of large language models (LLMs). It argues that LLMs should be viewed as models of language use rather than language cognition, emphasizing the need to incorporate sociolinguistic principles into the curation of training data. The paper highlights how sociolinguistic variation and the representation of different language varieties can enhance the societal value of LLMs. It also addresses the challenges of data contamination and evaluation practices in closed-source LLMs, advocating for a more transparent and socially responsible approach to AI development. The authors propose that integrating sociolinguistic theory, particularly the concept of lang... | |
| | | | |
ljvmiranda921.github.io
|
|
| | | | | Large language models showed promise on structured prediction tasks like named entity recognition and text categorization. But how well do they perform when ... | |
| | | | |
scottaaronson.blog
|
|
| | | Update (Dec. 17): Some of you might enjoy a 3-hour podcast I recently did with Lawrence Krauss, which was uploaded to YouTube just yesterday. The first hour is about my life and especially childhood (!); the second hour's about quantum computing; the third hour's about computational complexity, computability, and AI safety. I'm being attacked on... | ||
