News for hackers. Stuff that matters.

A comment on Building Reliable Systems Out of Unreliable Agents. Parent comment

Re:Building Reliable Systems Out of Unreliable Agents by maciejgryka on Tuesday April 09, 2024 @06:01PM (#39984775)
Ha, we tried that! Didn't make a noticeable difference in our benchmarks, even though I've heard the same sentiment in a bunch of places. I'm guessing whether this helps or not is task-dependent.
[ Parent | on HN ]
Re:Building Reliable Systems Out of Unreliable Agents by dudus on Tuesday April 09, 2024 @06:04PM (#39984793)
Agreed. I ran a few tests and observed similarly that threats didn't outperform other types of "incentives" I think it might some sort of urban legend in the community.

Or these prompts might cause wild variations based on the model and any study you do is basically useless for the near future as the models evolve by themselves.

[ Parent | on HN ]
Re:Building Reliable Systems Out of Unreliable Agents by maciejgryka on Tuesday April 09, 2024 @06:09PM (#39984829)
Yeah, the fact that different models might react differently to such tricks makes it hard. We're experimenting with Claude right now and I'm really hoping something like https://github.com/stanfordnlp/dspy can help here.
[ Parent | on HN ]
Re:Building Reliable Systems Out of Unreliable Agents by dollo_7 on Tuesday April 09, 2024 @06:04PM (#39984792)
I hoped it was too good to be just a joke. Still, I will try it on my eval set…
[ Parent | on HN ]
Re:Building Reliable Systems Out of Unreliable Agents by maciejgryka on Tuesday April 09, 2024 @06:07PM (#39984819)
I wouldn't be surprised to see it help, along with the "you'll get $200 if you answer this right" trick and a bunch of others :) They're definitely worth trying.
[ Parent | on HN ]