Model Alignment | Gelei Denghttps://geleideng.github.io/tags/model-alignment/Model AlignmentHugo Blox Builder (https://hugoblox.com)en-usTue, 02 Dec 2025 00:00:00 +0000https://geleideng.github.io/media/icon_hu7729264130191091259.pngModel Alignmenthttps://geleideng.github.io/tags/model-alignment/RSafe: Incentivizing Proactive Reasoning to Build Robust and Adaptive LLM Safeguardshttps://geleideng.github.io/publication/rsafe/Tue, 02 Dec 2025 00:00:00 +0000https://geleideng.github.io/publication/rsafe/<p>RSafe targets the brittleness of conventional guard models by making safety decisions through explicit reasoning over a user-specified policy, then reinforcing those reasoning traces to improve robustness on adversarial and out-of-distribution safety violations.</p>