Re:Show HN: Distill and serve small models with frontier qua...
by kfallah15 on Friday July 31, 2026 @12:54PM
(#49125672)
Great question. You need the simulator to be realistic enough that the optimizer cannot reward hack, but it does not need to be perfectly realistic.
[ Parent
| on HN ]
Re:Show HN: Distill and serve small models with frontier qua...
by SilenN on Friday July 31, 2026 @02:40PM
(#49127039)
^this
There's two ways to functionally measure this, reconstruction fidelity (which we're able to get to 0.7 - 0.95), and downstream performance (which agrees on the best and worst models, but can sometimes disagrees on the middle of the pack routing).