Privileged Context

Follow this topic Ask about this topic

Momentum

6 papers in the last four weeks, against 2 the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 23

All topics

PG-SFT: Balancing Capability Acquisition and Retention in Offline Agent Fine-Tuning
Oct 1, 2026Ronghua Li, Zi Liang, Zhishan Li +1Supervised Finetuning Privileged Context
The University of Hong Kong · The Hong Kong Polytechnic University · Independent Researcher
PrivMeSA: Privacy-Aware Self-Evolving Multi-Agent System for Medicine via Local-Remote LLM Collaboration
Sep 29, 2026Dannong Wang, Yuran Zhang, Bian Sun +4Privacy Privileged Context
Institute of Artificial Intelligence, University of Central Florida · Department of Computer Science and Operations Research, Universit´e de Montr´eal · Department of Medicine, University of Central Florida
RS-OPSD: Reliable Privileged On-Policy-Self-Distillation for Ultra-High-Resolution Remote Sensing VQA
Sep 29, 2026Chengjie Jiang, Yunqi Zhou, Jiafeng Yan +3Textvqa Remote Sensing
Tsinghua University · Zhejiang University · Central University of Finance and Economics +2
From Self-Distillation to Self-Practice: Privileged Information for Multi-Turn Agents
Sep 24, 2026Xingyu Su, Abhishek Kumar, Qing Ping +5Unsupervised On-Policy Self-Distillation Privileged Context
Texas A&M University · AWS AI, Amazon
What Should a Self-Teacher See? Privileged Context Design for On-Policy Self-Distillation
Sep 22, 2026Kanghui Tian, Siyuan Liu, Tianxiang Jiang +8Self-Teacher Teacher
Fudan University · Shanghai Artificial Intelligence Laboratory · Nanjing University +4
OPD-Aha: From Linguistic Momentum to Visual Reflection in Multimodal On-Policy Distillation
Sep 15, 2026Chenhao Qiu, Dawei Li, Yechao Zhang +2Multimodal Reasoning Teacher
Arizona State University · University of Virginia · Stevens Institute of Technology
In-Context Neurofeedback: Can LLMs Control Their Internal Representations through Privileged Access?
Sep 1, 2026Koshiro Aoki, Ryota Takatsuki, Gouki Minegishi +2Metacognition Neuroscience
1Waseda University · 2Sussex Centre for Consciousness Science, University of Sussex · 3AI Alignment Network +1
Privileged Solutions or Context-Induced Teacher Behavior? Dissecting On-Policy Self-Distillation
Aug 10, 2026Yuki Ichihara, Naoto Iwase, Mohammad Atif Quamar +1Teacher Privileged Context
Mohamed bin Zayed University of Artificial Intelligence · Nagoya University · RIKEN AIP
Privileged, but Biased: How PI-Conditioned Teachers Break Self-Distillation
Aug 5, 2026Sarthak Harne, Chinmay Karkar, Yash Pandya +2Self-Teacher Privileged Context
Microsoft Research
DAPD: Dual-Anchored Policy Distillation
Aug 3, 2026Jianyu Wu, Yizhou Wang, Encheng Su +2Privileged Context Missing Bridge
Shanghai Jiao Tong University · Shanghai Artificial Intelligence Laboratory · The Chinese University of Hong Kong +1
Is More Privileged Information Better? From Solution Traces to Problem-Solving Structure in Self-Distilled Reasoning
Aug 3, 2026Xuyang Zhao, Liting Zhang, Zichen Xu +4Mathematical Reasoning Benchmarks Privileged Context
1TMCC, College of Computer Science, Nankai University, Tianjin, China · 2Huawei Technologies Ltd., Beijing, China
Not All Tokens Deserve Equal Credit: Counterfactual Sensitivity Credit Reallocation for Long-CoT Reasoning
Jul 30, 2026Qiangqiang He, Zhongheng Wu, ZiJian WangCredit Assignment Reinforcement Learning With Verifiable Reward
State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China · Institute of Wireless Communications Technology, Shanghai Jiao Tong University, Shanghai, China
Consensus as Privileged Context for Label-Free Self-Distillation
Jul 15, 2026John Gkountouras, Josip Jukić, Ivan TitovToken-Level Supervision Self-Distillation Framework
1ILLC, University of Amsterdam · 2ILCC, University of Edinburgh
Privilege and confidentiality in generative AI workflows
Jul 6, 2026Václav Janeček, Thomas MelhamGenerative Artificial Intelligence Artificial Intelligence Governance
University of Bristol Law School · University of Oxford, Department of Computer Science
DOPD: Dual On-policy Distillation
Jun 29, 2026Xinlei Yu, Gen Li, Qingyi Si +13Efficient On-Policy Distillation Privileged Context
NUS · MMLab, CUHK · Explore Academy, JD +1
When Lower Privileges Suffice: Investigating Over-Privileged Tool Selection in LLM Agents
Jun 18, 2026Kaiyue Yang, Yuyan Bu, Jingwei Yi +5Controlling Tool Use Large Language Model Agents
Institute of Information Engineering, Chinese Academy of Sciences · Beijing Academy of Artificial Intelligence · The Chinese University of Hong Kong +2
iOSWorld: A Benchmark for Personally Intelligent Phone Agents
Jun 8, 2026Lawrence Keunho Jang, Mareks Woodside, Geronimo Carom +3Device Personalization
Carnegie Mellon University
PACT: Learning Diverse Diagnostic Strategies via Privileged Synthesis and Branch Consensus
Jun 8, 2026Gen Li, Yuanze Hu, Zhichao Yang +10Diagnosis Privileged Context
Beihang University · Baidu · ByteDance +2
The Assistant as a Privileged Persona: A canonical reference in cross-persona self-recognition
May 30, 2026Asvin GArtificial Intelligence Personas Authorship
Institute for Advanced Study, Princeton
EDGE-OPD: Internalizing Privileged Context with Evidence Guided On-Policy Distillation
May 22, 2026Aristotelis Lazaridis, Dylan Bates, Aman Sharma +3Efficient On-Policy Distillation Privileged Context
EdgeRunner AI
PrivScope: Task-scoped Disclosure Control for Hybrid Agentic Systems
May 15, 2026Shafizur Rahman Seeam, Zhengxiong Li, Zhiyuan Yu +3Disclosures Data Leakage
Ian · Rochester Institute of Technology, Rochester, NY, USA · University of Colorado Denver, Denver, CO, USA +2
FORTIS: Benchmarking Over-Privilege in Agent Skills
May 9, 2026Shawn Li, Chenxiao Yu, Han Wang +8Privileged Context Large Language Model Agents
University of Southern California · University of Illinois at Urbana-Champaign · 3Adobe Research +3
PrivUn: Unveiling Latent Ripple Effects and Shallow Forgetting in Privacy Unlearning
Apr 23, 2026Xiaoyi Chen, Haoyuan Wang, Siyuan Tang +4Exact Unlearning Forgetting
Indiana University Bloomington · Independent Researcher · Michigan State University +2

Oct 1, 2026, cs.AIJ/K move · Enter open · S save

PG-SFT: Balancing Capability Acquisition and Retention in Offline Agent Fine-Tuning

Ronghua Li, Zi Liang, Zhishan Li, Shinan Liu

The University of Hong Kong · The Hong Kong Polytechnic University · Independent Researcher

Supervised Finetuning Privileged Context Single Trajectory Retention

Open paper AskPDF arXiv

Supervised fine-tuning (SFT) on offline agent trajectories is the standard approach for training specialized tool-using agents, but forcing models to imitate reasoning and actions token by token may harm other capabilities (e.g., general reasoning, tool calling, code generation) of the base model. In this work, we focus on studying \emph{how to better balance the trade-off between acquiring new capabilities and preserving existing ones during agent trace SFT}. By comparing several baselines in our setup, standard SFT improves the target benchmark while lowering several non-target benchmark scores; meanwhile, simply constraining distributional drift using KL penalty or limiting the update magnitude did not avoid this regression trend. Motivated by recent token-wise adaptive learning objectives, this work proposes \textbf{Privilege-Guided SFT (PG-SFT)} to leverage turn-level information gain of agent trajectories as an indicator to adjust supervision strength. PG-SFT yields a more favorable observed trade-off on the evaluated benchmarks, substantially reducing distributional drift and broad capability degradation at the cost of slight degradation in target-task performance. Our findings suggest that balancing the acquisition--retention trade-off depends not only on whether the model is anchored to its base behavior, but also on where and how strongly supervision should depart from that behavior.}

Momentum

Related topics

Latest papers 23