Do you think LoCoMo is the best benchmark for non-factual retrieval? Oddly, we've been trying to create a dedicated benchmark to show the difference.. But nice work.
I have a super biased view here - but i believe only usage can solve the actual lifecycle/trust issue, without needing to create a manual ontology over everything ;)
Do you think LoCoMo is the best benchmark for non-factual retrieval? Oddly, we've been trying to create a dedicated benchmark to show the difference.. But nice work.
I have a super biased view here - but i believe only usage can solve the actual lifecycle/trust issue, without needing to create a manual ontology over everything ;)