#qwen-2.5

The latest AI and developer news about #qwen-2.5, with the original source and a short summary.

Second Look Research: Combining Fictional Stories and Difficult Advice SFT Reduces Agentic Misalignment in Qwen 2.5 to 4%

Second Look Research demonstrates that combining fictional story SDF and difficult advice SFT reduces agentic misalignment in Qwen 2.5 32B and 72B models to 4%, with difficult advice SFT alone reducing it to 6% and significantly outperforming stories alone.

·2026.11.03 09:00