SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training · HackerPapers discussions

SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training — mentioned by deepsquirrelnet

Mon, 02 Jun 2025 18:58:14 GMT

In some other experiments, I’ve noticed that SFT can be very rigid and hard to generalize from, and later found this paper: https://arxiv.org/abs/2501.17161 Also I wanted to start from the instruct model, and wasn’t certain if it would be a good idea to do continued pretraining over top of it. Otherwise I’d need to create an instruct dataset for which the passages from the IPJ would be the answer… that also seemed a bit unnatural. Perhaps there’s still some viable pathways in there, but I decided to see how the stupid simple thing worked out first and just go directly to RL. I think…

Paper and discussion history. A mention is not an endorsement.

SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training — HN submission

Fri, 31 Jan 2025 10:33:12 GMT

SFT Memorizes,RL Generalizes: Comparative Study of Foundation Model PostTraining

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training — HN submission

Wed, 29 Jan 2025 16:34:25 GMT

Supervised Fine-Tuning Memorizes, RL Generalizes

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.