cs.CL（2025-08-16）

📊 共 10 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

#	题目	一句话要点	标签	🔗
1	Capabilities of GPT-5 across critical domains: Is it the next breakthrough?	比较GPT-4与GPT-5在关键领域的能力，揭示其潜在突破	large language model multimodal
2	STEM: Efficient Relative Capability Evaluation of LLMs through Structured Transition Samples	提出STEM方法以高效评估大型语言模型的相对能力	large language model
3	J6: Jacobian-Driven Role Attribution for Multi-Objective Prompt Optimization in LLMs	提出J6以解决大型语言模型多目标优化问题	large language model
4	Mitigating Jailbreaks with Intent-Aware LLMs	提出Intent-FT以解决大语言模型的越狱攻击问题	large language model	✅
5	Mind the Generation Process: Fine-Grained Confidence Estimation During LLM Generation	提出FineCE以解决LLM生成过程中的置信度估计问题	large language model
6	CAMF: Collaborative Adversarial Multi-agent Framework for Machine Generated Text Detection	提出CAMF框架以解决机器生成文本检测问题	large language model
7	CORE: Measuring Multi-Agent LLM Interaction Quality under Game-Theoretic Pressures	提出CORE指标以量化多智能体LLM交互质量	large language model	✅

#	题目	一句话要点	标签
8	User-Assistant Bias in LLMs	提出用户助手偏见模型以优化多轮对话表现	DPO direct preference optimization large language model
9	Learning Wisdom from Errors: Promoting LLM's Continual Relation Learning through Exploiting Error Cases	提出基于指令的对比调优方法以解决持续关系学习中的错误案例问题	contrastive learning large language model instruction following
10	Optimizing Token Choice for Code Watermarking: An RL Approach	提出CodeTracer以解决代码水印嵌入的有效性问题	reinforcement learning policy learning