MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agent

ICLROral2026

Authors
Hongli Yu, Tinghong Chen, Jiangtao Feng, Jiangjie Chen, Weinan Dai, Qiying Yu, Ya-Qin Zhang, Wei-Ying Ma, Jingjing Liu, Mingxuan Wang, Hao Zhou
Affiliation
Tsinghua University
Venue
ICLR 2026
Track
Oral

TL;DR

We propose MemAgent, a novel agent workflow for long-text processing, demonstrating exceptional extrapolation and performance in large-scale tasks after RL Training.

Opening excerpt from the authors’ abstract. source

Read the paper

Topics

memory agent llm

← All ICLR 2026 Oral papers · Browse the whole archive