Sabela
@sabela.bsky.social
Developing infra for RL 👾🤖
Costa da Morte - Coruña - Zürich
Opinions are my own.
Costa da Morte - Coruña - Zürich
Opinions are my own.
Reposted by Sabela
I see a lot of discussion about this being an echo chamber. But, quite honestly, I think it's healthier to not be very aware of other people's views, than to always be exposed to the worst version of those views. X was not opening anyone's views, it was reinforcing how the others were wrong.
November 16, 2024 at 9:08 AM
I see a lot of discussion about this being an echo chamber. But, quite honestly, I think it's healthier to not be very aware of other people's views, than to always be exposed to the worst version of those views. X was not opening anyone's views, it was reinforcing how the others were wrong.
Our latest work on improving factuality for summarization models using RL is out! 🥳
arxiv.org/abs/2306.00186
arxiv.org/abs/2306.00186
Factually Consistent Summarization via Reinforcement Learning with...
Despite the seeming success of contemporary grounded text generation systems,
they often tend to generate factually inconsistent text with respect to their
input. This phenomenon is emphasized in...
arxiv.org
June 6, 2023 at 12:50 PM
Our latest work on improving factuality for summarization models using RL is out! 🥳
arxiv.org/abs/2306.00186
arxiv.org/abs/2306.00186
Reposted by Sabela
