method · first seen 9 Nov 2026
weighted supervised fine-tuning
Also: weighted SFT
Fine-tuning a model on examples weighted by scores, without reinforcement learning.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Let the Fixer Decide: Training a Code Localizer from the Fixer's LikelihoodPapers · 9 Nov 2026 · Khalil Hennara, Muhammad Hreden, Zeina Aldallal, Sara Chrouf, Safwan AlModhayan
Latest sources
- Let the Fixer Decide: Training a Code Localizer from the Fixer's LikelihoodPapers · 9 Nov 2026 · Khalil Hennara, Muhammad Hreden, Zeina Aldallal, Sara Chrouf, Safwan AlModhayan