来自EMNLP 2025，提出专利权利要求生成的多维评估基准，为LLM在专业文档生成领域提供新标尺。
Article URL: https://aclanthology.org/2025.emnlp-main.1564/ Comments URL: https://news.ycombinator.com/item?id=48683291 Points: 1 # Comments: 0
揭秘大模型内部单词生成的完整流水线，从原理到图表示，一篇学术深研论文。
arXiv:2606.08562v1 Announce Type: new Abstract: Transformer language models process input provided as subword fragments, but natural language semantic…
Transformer一致性训练机制，通过堆叠层间约束提升模型表现与稳定性。
arXiv:2606.05817v1 Announce Type: cross Abstract: Consistency training encourages models to behave similarly across different contexts, and has shown …
从最后层logits到逻辑，解锁LLM一致性的结构化知识推理新范式。
arXiv:2511.07910v2 Announce Type: replace Abstract: Large Language Models (LLMs) achieve excellent performance in natural language reasoning tasks thr…
大模型安全新方案：局部在策略蒸馏实现高效安全对齐，论文已提交EMNLP 2026。
arXiv:2606.02530v1 Announce Type: new Abstract: Aligning Large Language Models (LLMs) with human values often degrades their general capabilities, ter…
EMNLP 2025新研究：用多模态大模型提示与个性化偏好对齐，提升封面图像生成质量。
arXiv:2605.27374v1 Announce Type: new Abstract: Recent advances in multimodal large language models (MLLMs) and diffusion models (DMs) have opened new…
新研究提出受限路径上下文，探索如何控制LLM在知识图谱问答中的路径历史可见性，提升推理准确性。
arXiv:2605.26645v1 Announce Type: new Abstract: LLM-based knowledge-graph question answering (KGQA) delegates graph traversal to language models, turn…
EMNLP 2025 论文提出不确定性感知的效用锚点，让 AI 偏好优化更自适应、更鲁棒。
arXiv:2509.10515v1 Announce Type: cross Abstract: Offline preference optimization methods are efficient for large language models (LLMs) alignment. Di…