Closed imoneoi closed 4 months ago
Thank you, looking forward to update this data point.
I note by the way that the paper talks of 'open-source' models where actually most of these are at best 'open weight' — nothing about the source of most of the tested models is actually open.
@imoneoi is there a canonical URL or a DOI? Openreview.net review forum IDs are not guaranteed to be durable links. (If the arxiv.org submission is updated with the accepted version and includes a note like 'accepted at ICLR', we would prefer that.)
OpenChat paper is recently accepted by ICLR 2024 https://openreview.net/forum?id=AOJyfhWYHf