cs.CL（2025-12-14）

📊 共 14 篇论文

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (12) 支柱二：RL算法与架构 (RL & Architecture) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (12 篇)

#	题目	一句话要点	标签	🔗	⭐
1	Persistent Personas? Role-Playing, Instruction Following, and Safety in Extended Interactions	提出长程对话评估协议，揭示LLM角色扮演中身份一致性衰退问题	large language model instruction following
2	Auto-Tuning Safety Guardrails for Black-Box Large Language Models	针对黑盒大语言模型，提出基于超参数优化的安全护栏自动调优方法	large language model
3	Human-Inspired Learning for Large Language Models via Obvious Record and Maximum-Entropy Method Discovery	提出一种受人类启发的大语言模型学习框架，解决罕见场景下的泛化问题。	large language model
4	ERA-IT: Aligning Semantic Models with Revealed Economic Preference for Real-Time and Explainable Patent Valuation	提出ERA-IT框架，通过经济偏好对齐语义模型，实现实时且可解释的专利估值。	large language model chain-of-thought
5	HyperEdit: Unlocking Instruction-based Text Editing in LLMs via Hypernetworks	HyperEdit：通过超网络解锁LLM中基于指令的文本编辑能力	large language model
6	Does Tone Change the Answer? Evaluating Prompt Politeness Effects on Modern LLMs: GPT, Gemini, LLaMA	评估提示语礼貌程度对大型语言模型的影响：GPT、Gemini、LLaMA	large language model
7	State over Tokens: Characterizing the Role of Reasoning Tokens	提出State over Tokens框架，将LLM推理token视为外部计算状态而非语言叙事。	large language model
8	Fine-Tuning Causal LLMs for Text Classification: Embedding-Based vs. Instruction-Based Approaches	探索高效微调因果LLM文本分类方法：嵌入式 vs. 指令式	large language model
9	DeliberationBench: When Do More Voices Hurt? A Controlled Study of Multi-LLM Deliberation Protocols	DeliberationBench揭示多LLM协商一致协议效果不佳，优于单模型最佳选择的假设被证伪	large language model
10	Understanding Syllogistic Reasoning in LLMs from Formal and Natural Language Perspectives	研究大型语言模型在形式语言和自然语言视角下的三段论推理能力	large language model
11	AnimatedLLM: Explaining LLMs with Interactive Visualizations	AnimatedLLM：通过交互式可视化解释大型语言模型	large language model
12	LexRel: Benchmarking Legal Relation Extraction for Chinese Civil Cases	LexRel：构建中文民事案件法律关系抽取基准，揭示LLM局限性	large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
13	CoDA: A Context-Decoupled Hierarchical Agent with Reinforcement Learning	提出CoDA：一种解耦上下文的分层强化学习Agent，解决LLM Agent中的上下文爆炸问题。	reinforcement learning large language model
14	Coupled Variational Reinforcement Learning for Language Model General Reasoning	提出耦合变分强化学习CoVRL，提升语言模型通用推理能力	reinforcement learning

⬅️ 返回 cs.CL 首页 · 🏠 返回主页