Update README.md

README.md

@@ -26,7 +26,7 @@ tags:
       alt="ContextPilot Live Demo"
     />
   </a>
-  <a href="">
+  <a href="https://arxiv.org/abs/2608.28476">
     <img
       src="https://img.shields.io/badge/ContextPilot-Paper-red?logo=arxiv&logoColor=red"
       alt="Paper"
@@ -40,7 +40,7 @@ tags:
   </a>
 </p>
 
-ContextPilot-14B is the Qwen3-14B checkpoint of **ContextPilot**, a proactive context-management framework for long-horizon language-model agents. It teaches agents to plan, maintain long-term memory, and offload less useful context while they continue reasoning and using tools. For more details, see our [paper]() and [code repository](https://github.com/Tencent/ContextPilot).
+ContextPilot-14B is the Qwen3-14B checkpoint of **ContextPilot**, a proactive context-management framework for long-horizon language-model agents. It teaches agents to plan, maintain long-term memory, and offload less useful context while they continue reasoning and using tools. For more details, see our [paper](https://arxiv.org/abs/2608.28476) and [code repository](https://github.com/Tencent/ContextPilot).
 
 ![ContextPilot overview](assets/contextpilot_overview.png)
 
@@ -80,4 +80,21 @@ This checkpoint is intended for research on proactive context management, long-h
 
 ## Citation
 
-Citation details will be added with the accompanying paper.
+```
+@inproceedings{pan-etal-2026-contextpilot,
+    title = "ContextPilot: Teaching Agents for Proactive Context Management via Fine-grained RL",
+    author = "Pan, Zhuoshi  and
+      Pei, Qizhi  and
+      Lu, Junru  and
+      Lin, Honglin  and
+      Zhao, H. Vicky  and
+      Yin, Di  and
+      Sun, Xing",
+    booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
+    month = nov,
+    year = "2026",
+    address = "Budapest, Hungary",
+    publisher = "Association for Computational Linguistics",
+    abstract = "Long-horizon agentic tasks require large language models (LLMs) to iteratively retrieve, integrate, and maintain dispersed information across multi-turn interactions, but preserving all interaction histories leads to a continuously growing working context. Recent proactive context management methods allow models to edit their own working context with specialized tools, yet they still face three key limitations: (1) a limited toolset restricted to search, deletion, and summarization, with no support for global planning, long-term memory, and adaptive compression; (2) inefficient exploration that treats context management actions uniformly despite their heterogeneous impacts on final outcomes; and (3) coarse-grained credit assignment that assigns the final trajectory-level reward to all intermediate context editing actions during RL. To bridge these gaps, we introduce ContextPilot, a proactive context management framework for long-horizon agentic reasoning. Our approach systematically augments the toolset with planning, long-term memory, and soft context offloading tools. We further propose an RL method tailored for context management, which uses context and entropy variation to identify critical editing decisions for branch sampling and estimates action-level advantages from all branched trajectories that pass through the corresponding context editing action. Experiments on long-context QA and deep search tasks show that ContextPilot achieves stronger performance with a more compact working context, consistently outperforming existing baselines across various base models and benchmarks. Code is available at \url{https://github.com/Tencent/ContextPilot}."
+}
+```