arXiv cs.AI发布于 09/04 12:00
Caught in the Story: Narrative Captivity in Multi-turn LLMs Conversation
（翻译）陷入故事：多轮LLM对话中的叙事俘获
Abstract page for arXiv paper 2609.03407: Caught in the Story: Narrative Captivity in Multi-turn LLMs Conversation
arXiv cs.AI发布于 09/12 12:00
KuaiRP Series Role-playing Models Technical Report
（翻译）KuaiRP 系列角色扮演模型技术报告
Abstract page for arXiv paper 2609.11127: KuaiRP Series Role-playing Models Technical Report
Google AI Blog发布于 09/16 00:00
AI for everyone in every language
（翻译）让 AI 覆盖每一种语言和每一个人
We’re moving beyond traditional text translation to build models that understand the world’s rich, living languages exactly as they are expressed.
arXiv cs.AI发布于 09/23 12:00
Clarification Is Not Correction: LLMs Fail to Let Go
（翻译）澄清不是纠正：大语言模型无法放手
Abstract page for arXiv paper 2609.25337: Clarification Is Not Correction: LLMs Fail to Let Go
arXiv cs.AI发布于 09/25 12:00
Pistis Technical Report
（翻译）Pistis 技术报告
Abstract page for arXiv paper 2609.28554: Pistis Technical Report
arXiv cs.AI发布于 09/25 12:00
TWIST: A Proposed Benchmark for Intervention Quality in Conversational Memory, with a Human-Validated Draft-Alignment
（翻译）TWIST：面向对话记忆干预质量的拟议基准，附带经人工验证的草稿对齐
Abstract page for arXiv paper 2609.28575: TWIST: A Proposed Benchmark for Intervention Quality in Conversational Memory, with a Human-Validated Draft-Alignment
arXiv cs.AI发布于 09/25 12:00
Beyond Surface Style: Aligning Multi-Turn User Simulators with Behavioral Consistency
（翻译）超越表面风格：使多轮用户模拟器与行为一致性对齐
Abstract page for arXiv paper 2609.28690: Beyond Surface Style: Aligning Multi-Turn User Simulators with Behavioral Consistency
arXiv cs.AI发布于 09/28 12:00
Do LLMs Understand Context? A Knowledge Graph-Based Evaluation Framework
（翻译）大语言模型理解上下文吗？一个基于知识图谱的评估框架
Abstract page for arXiv paper 2609.30484: Do LLMs Understand Context? A Knowledge Graph-Based Evaluation Framework
arXiv cs.AI发布于 09/28 12:00
Audio LLMs Know When They Can't Hear You
（翻译）音频大模型知道自身何时听不清
Abstract page for arXiv paper 2609.30625: Audio LLMs Know When They Can't Hear You
arXiv cs.AI发布于 10/05 12:00
Fast Models, Slow Evidence: A Paired and Self-Audited Evaluation of System-1 Decision Models for LLM Agent Harnesses
（翻译）快速模型，缓慢证据：面向 LLM Agent Harnesses 的 System-1 决策模型配对与自审计评估
Abstract page for arXiv paper 2610.02267: Fast Models, Slow Evidence: A Paired and Self-Audited Evaluation of System-1 Decision Models for LLM Agent Harnesses
arXiv cs.AI发布于 10/08 12:00
How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis
（翻译）端侧语言模型安全性有多脆弱？用于稀疏故障分析的安全关键参数定位
Abstract page for arXiv paper 2610.09000: How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis
arXiv cs.AI发布于 10/08 12:00
Sigma-Hunter: A Domain-Specific Language Model for Threat Hunting and Detection Engineering
（翻译）Sigma-Hunter：用于威胁狩猎和检测工程的领域特定语言模型
Abstract page for arXiv paper 2610.09007: Sigma-Hunter: A Domain-Specific Language Model for Threat Hunting and Detection Engineering
arXiv cs.AI发布于 10/08 12:00
Enabling Dynamic Computation in Looped LMs
（翻译）在循环式语言模型中启用动态计算
Abstract page for arXiv paper 2610.09013: Enabling Dynamic Computation in Looped LMs
arXiv cs.AI发布于 10/09 12:00
Beyond Imitation: A Framework and Benchmark for LLM-Assisted Peer Review
（翻译）超越模仿：LLM 辅助同行评审的框架与基准
Abstract page for arXiv paper 2610.11087: Beyond Imitation: A Framework and Benchmark for LLM-Assisted Peer Review