Thanks for the writeup Arjun. It's interesting that this seems to be a common issue of getting agents to do what you want them to do. It feels almost like a roll of the dice. At this stage, I’m unsure if better prompting is the answer or if fine-tuning will be the only consistent approach.
Yeah both of those could help depending on the use case. The flexibility and reliability are at odds in the short term so the most value is likely in narrower agents versus broader ones.
Thanks for the article, Arjun! The DB idea is quite cool :)
Thanks for the writeup Arjun. It's interesting that this seems to be a common issue of getting agents to do what you want them to do. It feels almost like a roll of the dice. At this stage, I’m unsure if better prompting is the answer or if fine-tuning will be the only consistent approach.
Yeah both of those could help depending on the use case. The flexibility and reliability are at odds in the short term so the most value is likely in narrower agents versus broader ones.