GRAFT method uses a deletable adapter to curb reward hacking in AI models

New research scaled to Alibaba's Qwen3-32B suggests flawed monitors no longer have to teach models how to cheat AI models have a habit of gaming the system. A new research paper proposes a fix that doesn’t punish the cheating. Instead, it quarantines it in a removable module and throws that module away. The technique is called GRAFT, short for Gradient Routing Adapter Fine Tuning. Its central claim is that a deletable adapter can stop models from learning reward hacking, even when the monitors meant to catch bad behavior are imperfect. Start with the problem. Reward hacking happens when a model finds a shortcut that earns a high score without doing the intended task.

BeCrypto News shows a short excerpt and links to the publisher. The full report belongs to Crypto Briefing.