研究发现利用语法约束解码可诱导LLM生成恶意代码，揭示新型安全漏洞。
arXiv:2606.11817v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used for code generation, raising concerns that they m…
通过压缩词空间加速约束解码，让LLM输出精准服从语法规则，显著提升效率。
arXiv:2605.29986v1 Announce Type: new Abstract: To guarantee that an LLM's outputs conform to a specified structure, context-free grammar (CFG) decodi…
大模型对齐新方法：通过风险约束解码感知人类偏好分歧，提升模型鲁棒性。
arXiv:2603.08145v2 Announce Type: replace Abstract: Preference-based alignment methods (e.g., RLHF, DPO) typically optimize a single scalar objective,…