@OxfordAI
Our paper, “VisoGender: A dataset for benchmarking gender bias in image-text pronoun resolution” has been accepted into #NEURIPS2023. VisoGender is an evaluative tool designed to flag a model’s bias prior to deploymentView quoted post on X
arxiv.org/abs/2306.12424
youtu.be/swe3JvoHUBo?si… 1/n
Interested in AI safety and large language models? You should join @sleepinyourhat talk at AISHED.View quoted post on X
www.facebook.com AGI Governance: Why and How? (Sam Clarke, GovAI) in Oxford, United... Networking event in Oxford, United Kingdom by Oxford Artificial Intelligence Society on Tuesday, November 15 2022 AI development could have huge benefits, but also huge risks - so how can we mitigate possible catastrophic risk from advanced AI systems?
Join us next Tuesday to hear Sam Clarke's (@GovAI_ ) introduction to AGI governance fb.me/e/3CjyUdbX7View quoted post on X
